SPSS Analysis Unit 5

njosh

unit5homework_0.pdf

Home >Mathematics homework help >SPSS Analysis Unit 5

What do you want to do?

How many variables?

What level of data?

Central tendency

Describe

Median

Make inferences

Univariate Bivariate Multivariate

Nominal Ordinal Interval/

Ratio

Central tendency

Mode

Central tendency

Mean

Form

Skew

Form

Skew

Kurtosis Kurtosis Kurtosis

Form

Skew

Dispersion Dispersion Dispersion

Variance, Standard deviation

Average Absolute deviation

Range, Index of

dispersion

16304_CH06_Walker.indd 146 7/12/12 10:02:10 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Learning Objectives

Understand the number of modes, skewness, and kurtosis as they relate to ■■

explaining a data set.

Explain the difference between the mode and the number of modes.■■

Interpret the values of skewness and kurtosis as they relate to univariate ■■

analysis.

Discuss the importance of the normal curve in statistics.■■

Describe the properties of the normal curve.■■

The final univariate descriptive statistic, the form of the distribution, ties together the central tendency and dispersion of the data. Three characteristics make up the form of the distribution: the number of modes, the symmetry, and the kurtosis. In addressing the form of a distribution, a polygon can generally be used to represent these charac- teristics visually.

1 Moments of a Distribution6- In some statistics books and other places, distributions and the form of distributions are referred to in terms of the moments of the distribution. There are four moments that are considered important to a distribution. Moments are calculated as follows:

S1X 2 X2 i N

where 1X 2 X2 represents the deviations from the mean (as has been the case in Chap- ters 4 and 5), N is the total number of cases in the distribution, and i is the moment being calculated.

Since the sum of the deviations around the mean is always zero, the first moment is always zero. If X is taken to the second power in the formula above, you can see that this is the formula for the variance—thus, the second moment is the variance. The third moment is usually associated with the skew of the distribution, although the exact formula is to divide the formula for the third moment by the variance to the power of 1.5. Similarly, the kurtosis of a distribution is associated with the fourth

Chapter 6

The Form of a Distribution

147

16304_CH06_Walker.indd 147 7/12/12 10:02:10 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

moment, although the exact formula is to divide the formula for the fourth moment by the variance squared. The mean and variance were discussed in “Measures of Central Tendency” and “Measures of Dispersion” (Chapters 4 and 5). The skew and kurtosis of a distribution are discussed in this chapter, together with the third measure of the form of a distribution: the number of modes.

2 Number of Modes6- The first measure of the form of a distribution is the number of modes. The number of modes is important to higher-order analyses because it is indicative of the normal- ity of the distribution. To use many bivariate and multivariate statistical procedures, a unimodal distribution is preferred.

In determining the number of modes, a slight deviation from determining the mode may be necessary. Recall from the discussion of central tendency that it is com- mon to count only the highest frequency in a distribution as the mode, even though some people argue that all peaks of a distribution should be considered. For determin- ing the number of modes in an analysis of form, it may be more beneficial to look at peaks rather than to find the one, highest value. Consider, for example, the distribution in Figure 6-1. Even though there is only one highest value, there are three peaks in the distribution. These peaks may make the data unsuitable for certain statistical pro- cedures unless transformations are made. In this distribution, all three modes should probably be counted in evaluating the form of the distribution even though the mode is actually only 4.

0 1 2 3 4 5 6 7 8 109

Figure 6-1 Polymodal Distribution

148 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 148 7/12/12 10:02:10 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

3 Skewness6- The next characteristic of the form of the distribution is the degree of symmetry (skewness) of the distribution. This measure of the form of a distribution has three cat- egories: symmetrical, positively skewed, and negatively skewed. A fully symmetrical distribution has mirror-image sides such that the distribution could be split at the mean and the sides folded over each other for a perfect match. In Figure 6-2, it is easy to see the symmetry in the distribution. This is the histogram from Figure 4-7. The frequen- cies displayed in this distribution are very balanced: categories 1 and 7 have the same frequency, as do 2 and 6, and 3 and 5. Category 4 has the highest frequency level. It is easy to see that this distribution could be folded in half and the two sides would match perfectly. This distribution is therefore, a perfectly symmetrical distribution. In actual research, however, it is not common to see a perfectly symmetrical distribution. More typically, the distribution will either be only close to symmetrical or not at all symmetrical.

It should be noted here that the number of modes does not necessarily affect the skew of the distribution. A distribution that is bimodal could still be cut in half where the distribution mirrors itself. The only difference in this case is that the mode and other measures of central tendency would not be the same.

1 2 3 4 5 6 7

Figure 6-2 Histogram and Normal Curve for a Symmetrical Distribution

1496-3 Skewness

16304_CH06_Walker.indd 149 7/12/12 10:02:10 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Analysis of Skew

If a distribution is such that one side is different from the other, it is said to be skewed. In skewed distributions, there is no point that can be drawn in the polygon where it could be divided into two similar parts. If the point of the curve is to the left of the graph, it is said to be negatively skewed (the tail of the graph points to the negative end of the scale, smaller positive numbers). In Figure 6-3, “children” is an example of a nega- tively skewed distribution. Here, the point of the curve is toward category 1 or the left of the graph. If the point of the curve is to the right of the graph, it is said to be positively skewed (the tail of the graph points toward the positive end of the scale, larger positive numbers). In Figure 6-3, “gun-wher” is an example of a positively skewed distribution. Here, the point of the curve points toward category 12.5 or the right of the graph.

SPSS provides measures of skew in frequency output. A value of 0 means there is no skew to the data. Skew values of zero are almost never obtained, however, and a distribution is considered symmetrical if the skew value in SPSS is between 21 and 1.1 A distribution is generally considered skewed if it has a skew greater than +1.00 or less than (a greater negative number) 21.00. The magnitude of the number will represent the degree of skew. When conducting research, it is desirable to obtain a distribution that has a skew as close to zero as possible. If the skew is outside +1 to 21, the distribution may be too skewed to work with, and efforts should be made to get the distribution closer to normal. This is done through transformations, which is addressed in the discussion on regression.

The frequency distribution that has been used with the other univariate measures is shown in Table 6-1. Here the value of skew is 20.477, which means that the dis- tribution is not perfectly symmetrical but that it exhibits an acceptable level of skew (it is within the acceptable range of 0 to 21.00). There is some negative skew to this distribution, as exhibited by the negative value, but it is not enough to warrant addi- tional analyses or give cause for concern. If this value had been less than 21.00 (e.g., 22.77), it might have been necessary to transform the distribution.

CHILDREN

2.001.501.00

400

300

200

Std. Dev. = 0.22 Mean = 1.95 N = 324.00

100

GUN_WHER

12.510.07.55.02.50.0

Std. Dev. = 1.85 Mean = 2.2 N = 47.00

Figure 6-3 Negatively and Positively Skewed Distributions

150 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 150 7/12/12 10:02:11 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Although quantitative measures of skewness and kurtosis will almost always be available when conducting actual research, an estimate of the skew of a distribution can be made even without a skew calculation. If the mean and the median are different, the distribution is at least somewhat skewed, although it is not possible to tell if it is beyond +1 or 21. Additionally, the skew is in the direction of the mean. For example, if the distribution is positively skewed, the mean will be larger than the median, but if the skew is negative, the mean will be smaller than the median. It should also be noted that the mode is generally on the opposite side of the median from the mean in skewed distributions. This is not always the case, however, and should not be treated as a rule. MacGillivray (1981) discusses the conditions under which each of these examples would fall.

What is your highest level of education?

Value Label Value Frequency Percent Valid

Percent Cumulative

Percent

Less than High School 1 16 4.6 4.8 4.8

GED 2 59 17.0 17.6 22.3

High School Graduate 3 8 2.3 2.4 24.7

Some College 4 117 33.7 34.8 59.5

College Graduate 5 72 20.7 21.4 81.0

Post Graduate 6 64 18.4 19.0 100.0

Missing 11 3.2

Total 347 100.0 100.00

N Valid 336

Missing 11

Mean 4.08

Median 4.00

Mode 4

Std. Deviation 1.460

Variance 2.131

Skewness 2.477

Std. Error of Skewness .133

Kurtosis 2.705

Std. Error of Kurtosis .265

Table 6-1 SPSS Output of Measures of Form

1516-3 Skewness

16304_CH06_Walker.indd 151 7/12/12 10:02:11 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

4 Kurtosis6- The last characteristic of the form of a distribution is the kurtosis. For kurtosis, think again of stacking blocks (or beer cans) on top of each other to represent the frequency of the categories in a histogram. The kurtosis is the extent to which cases are piled up around the measure of central tendency or in the tails of the distribution. If most of the values in the distribution are very close to the measure of central tendency, the dis- tribution is said to be leptokurtic (as shown on the left in Figure 6-4). If most of the values in the distribution are out in the tails, the distribution is said to be platykurtic, as shown on the right in Figure 6-4. If the values in the distribution are such that they represent a distribution such as that shown in Figure 6-2, the distribution is said to be mesokurtic, as shown in the center of Figure 6-4. It is desirable to have a mesokurtic distribution in research; otherwise, the data may have to be transformed.

The shape of these curves also offers an opportunity to talk about variance and standard deviation. As discussed in “Measures of Dispersion” (Chapter 5), the vari- ance and standard deviation dictate the shape of the distribution. In a leptokurtic dis- tribution, the variance and standard deviation would be smaller than in a mesokurtic distribution. The variance and standard deviation of a platykurtic distribution would be larger than either a mesokurtic or leptokurtic distribution. This is one application of the variance and standard deviation. A more expanded discussion of this application is presented later in the section on the normal curve.

Analysis of Kurtosis

In SPSS, kurtosis is measured in the same way as skew. A value between +1 and 21 represents a mesokurtic distribution. Positive numbers greater than 1 represent lepto- kurtic curves. Negative numbers less than 21 (a greater negative number) represent platykurtic curves. As with skew, it is desirable to get the kurtosis as close as possible

1 2 3 4 5 6 71 2 3 4 5 6 7 1 2 3 4 5 6 7

Figure 6-4 Leptokurtic, Mesokurtic, and Platykurtic Distributions

152 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 152 7/12/12 10:02:11 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

to zero, using transformations if necessary. Examining the kurtosis value in Table 6-1 shows that the distribution is mesokurtic because the value (20.705) is between 21.00 and 0. If this value had been 21.705, the distribution would have been platykurtic.

5 The Importance of Skew and Kurtosis6- It is important to know the skew and kurtosis because some statistical procedures do not work well with skewed data or data that is not mesokurtic. If data in a research project is found to be skewed or kurtose, it may be necessary to transform the data. Initially, you must remember two things about transformation. First, if the data is not within accept- able tolerances for skew and kurtosis, the data will need to be transformed prior to using some statistical procedures. Second, after making transformations, recheck both the skew and kurtosis. Transforming the data may bring one of these within acceptable tol- erances but may make the other unacceptable. If that happens, you will need to choose another transformation. You should then recheck the skew and kurtosis again. This process should continue until you reach a point where both the skew and kurtosis are acceptable. If it is not possible to get both the skew and kurtosis in an acceptable range, you may need to consider using a different analysis procedure that is not susceptible to nonnormal curves.

6 Design of the Normal Curve6- Extending the concepts of the frequency distribution, graphical representation of data, and measures of central tendency, dispersion, and form brings us to the point of dis- cussing a key concept in statistical analysis, the normal curve. At this point, you should not be concerned with applications of the normal curve; that is covered in more detail in the chapters of the book on inferential analysis. The purpose of the present discussion is to introduce the properties of the normal curve.

An introduction to the normal curve is included in descriptive analyses rather than inferential analyses for two reasons. First, the normal curve can be used to provide an interpretation of the variance and standard deviation. Second, the normal curve is important to a number of statistical procedures that will be discussed before reaching information on inferential statistical procedures.

An example of relatively normally distributed data can be shown in grades in a course (see Figure 6-5). Say that most people taking the course score a C on the first test. This would be the modal grade (the top part of the curve). There are those who receive high A’s, but there would be only a few of these; they would be at the positive end of the curve. There are also those who receive very low F’s, but these are also few; they would be at the negative end of the curve. Most people would be in between these two extremes, with more people making scores around C’s than other grades, and

1536-6 Design of the Normal Curve

16304_CH06_Walker.indd 153 7/12/12 10:02:11 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

more people making B’s and D’s than A’s and F’s. For the sake of argument, though, say that no one received a 100 and there were a few people who did not take the test. Therefore, the ends of the tails will never completely touch the baseline.

This type of data represents a special form of distribution called the normal curve. This type of curve or distribution is very much like those that have been used in this chapter and at the end of Chapter 4. A normal curve is special because it has certain characteristics. First, a normal curve is symmetrical in that it can be folded in half, and both sides would be exactly the same (as in Figure 6-2, where the frequencies of category 1 and 3 are the same as those of category 5 and 7, respectively). Note, though, that this does not mean this curve cannot be kurtose. Some symmetrical distributions are leptokurtic or platykurtic. This is shown in Figure 6-4, where each of those dis- tributions was symmetrical even if it was kurtose. Also, a normal curve is unimodal; there is one, and only one, peak. This peak is at the maximum frequency of the data distribution, so that the mean, median, and mode all have the same value. The normal curve shown in Figure 6-5 has only one mode. From the peak, the tails of a normal curve fall off on both ends and extend to infinity, always getting closer to the baseline but never touching it. This is shown in Figure 6-5, where the bottom part of the curve straightens out and runs relatively parallel to the X axis. You may say this makes no sense; all distributions have an end, so why would the normal curve not have an end? The answer lies in the scientific process. Take the example of computers. Less than 20 years ago, scientists and engineers thought they had achieved the ultimate when they were able to reach 640K of random access memory (RAM) in a computer. They felt that this was the maximum that could be achieved and all that anyone would ever need. To them, the distribution limits were set. We now know, of course, that 640K was only

Figure 6-5 Normal Curve

154 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 154 7/12/12 10:02:11 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

the beginning and that computers are far beyond that now. It would have been foolish, then, to have the curve touch the line at 640K; it should not touch the line because we do not know what will come in the future. A final characteristic of the normal curve that merits discussion is that the area under a normal curve is always the same, regard- less of the data set. The area under the normal curve is 1.00, or 100% of all values in the distribution. This is extremely important in the section of this book concerning inferential analyses because of its importance in estimating the placement of a sample distribution within a population or another sample.

The area under the normal curve also offers the opportunity to put the variance and standard deviation into practice. Say, for example, that a researcher was examin- ing the time prisoners were out on parole before they committed another crime or returned to prison on a technical violation. If the time each parolee took before being reincarcerated was plotted, it might look as in Figure 6-6. There were a few people who returned to prison right away, most of the parolees who returned to prison did it within two to four years, and some took longer. Some had not recidivated at the time of the research, so the end of the distribution is open.

An analysis of the central tendency would put the mean of this distribution at 36 months, which is represented by the vertical line. This is good information: the aver- age length of time for parolees to be reincarcerated is three years. It is obvious from this distribution, however, that not all of the parolees were reincarcerated at the same time. The span of time runs from a couple of months to more than five years. To get a

Months

12 24 36 48 60

F re

q u

e n

c y

Figure 6-6 Distribution of Time to Reincarceration for Parolees

1556-6 Design of the Normal Curve

16304_CH06_Walker.indd 155 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

more accurate picture of the distribution of parolees, we might want to know, on aver- age, how far each of them is from the mean.

To calculate this, the procedure would be to take each person (each dot in Figure 6-6) and determine how far from the mean it is. This could be completed by measuring the distance with a ruler, but because this is a numeric scale, it could also be completed

by subtracting each value from the mean. This operation would be noted as X 2 X.

Summing all of these values, represented by S1X 2 X2, would provide us with the total of the distance from each value to the mean. But remember that the sum of each value subtracted from the mean is zero, so that does not help us. A solution to this is

to square each value before summing it. This calculation, S1X 2 X22, will give a posi- tive value. Knowing the total distance between the mean and each value is good, but a simpler value would be the average of the distance from each value to the mean. The calculation for this procedure is:

S1X 2 X22 N

The only problem with this calculation is that the value that would be obtained is not on the same scale as the original data. To return this value to the original scale, take the square root of the value. Of course, you recognize that this is the procedure for calculating the variance and, ultimately, the standard deviation for the distribution. Although this procedure can be completed for any distribution, it is particularly impor- tant for a normal curve because of what the standard deviation represents.

Because the area under the normal curve is always 100% and because standard deviations represent standard distances in relation to the mean of the distribution, stan- dard deviations can be used to calculate the area under the normal curve for particular values. For example, between the mean and 1 standard deviation under a normal curve lies 34.13% of all the data. Between the mean and 2 standard deviations is 47.72% of all the data. Between the mean and 3 standard deviations is 49.87% of all the data. That covers 49.87% of the possible 50% of one half of the curve. Since the normal curve is symmetrical, the same values can be obtained whether counting from the left or right of the mean. This also means that the values could be doubled from one side and the area under the whole normal curve could be determined for values plus or minus a certain number of standard deviations from the mean. If the figures above are doubled, the result would be 68.26% of the data between 21 and 1 standard deviation, 95.44% of the data between 22 and 2 standard deviations, and 99.74% of the data between 23 and 3 standard deviations.

This is very useful to know because it is now possible to determine what percent- age of the scores lie between the mean and any standard deviation within the distribu- tion. The problem is that researchers are not always interested in determining values (or determining the area under a normal curve) only for values directly on a particular

156 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 156 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

standard deviation. Often, calculations need to be made that allow any value in the distribution to be converted to standard deviation units such that the area under the normal curve can be calculated for that value. The procedure for doing this is to calcu- late a Z score for that value. Z scores convert any value in the distribution to standard deviation units. The formula for calculating a Z score is

Z 5 X 2 X

where X is the number to be converted to a Z score, X the mean of the distribution, and s is the standard deviation of the distribution.

The process of converting values to Z scores is simply a method of deriving stan- dard scores. To be able to make comparisons between things, it is necessary to set a standard on which both items may be measured. If measuring distance, we could use feet or meters; weight could be in pounds or kilograms. But what about social data such as crime? Crime is generally discussed in terms of crime rates. Calculating crime rates is simply a process of standardizing crime in terms of the population of a city— taking numbers that may be difficult to compare between different things and making them comparable. The same can be done with the mean, standard deviation, and nor- mal curve through Z scores. This standardizes scores based on the normal curve.

What this means is that if a researcher knows the mean and standard deviation of a population, Z scores can be used to calculate the distance between any value and the mean. Using the area under the normal curve allows inferential analyses (see the chapters in this book on inferential analyses) to be used to determine the chances that any number in a distribution will be in a sample drawn from that population.

The process of calculating a Z score, determining the area under the normal curve between that value and the mean, and examining how that score relates to the mean, is rather simple. First, subtract the raw score from the mean. This determines whether the number is above or below the mean: negative numbers are below the mean, positive numbers above the mean. Then divide the number by the standard deviation to deter- mine how many standard deviation units the number is above or below the mean. The answer obtained from this calculation is the score’s deviation from the mean in stan- dard units (the Z score in standard deviation units). Note that you may get a negative number; all this means is that the Z score is to the left of the mean. It does not affect the calculations at all and can be dropped in the remainder of the procedure.

The next step is to take this number and turn to Table B-1 in Appendix B, “Sta- tistical Tables.” The row to use corresponds to this number in column a of the table. The columns to use within this row depend on what area under the normal curve is being examined. If you are looking for the area between the score and the mean, look to column b in the table; if examining the area beyond that score, which is farther into the tail of the distribution, use column c.

1576-6 Design of the Normal Curve

16304_CH06_Walker.indd 157 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Suppose that the number of complaints against a police department averaged 12 complaints a month with a standard deviation of 1.5. Then suppose that the chief wants you to determine what percentage of the scores fell between the mean and the current month’s score of 14. This would require determining the area between the mean and 14. To do this, first calculate a Z score, as shown below:

Z 5 X 2 X

5 14 2 12

1.5

5 2

1.5

5 1.33

Then look at column a in Table B-1 in Appendix B, “Statistical Tables,” and find 1.33. You are looking for the area between the score and the mean, so you would use col- umn b to get the percentage of the area under the normal curve that falls between the mean and 14. This value is 0.4082, so the area in this distribution between 12 and 14 is 40.82%. What does this mean? It means that almost 41% of the months in the current year had fewer complaints recorded than those in the current month.

But what if the chief wanted to know how many months had more complaints recorded than the current month? To determine the percentage of scores greater than 14, you would calculate the Z score exactly as before. That number, 1.33, is then found in column a of Table B-1 in Appendix B, “Statistical Tables.” Since you are trying to determine the number of scores greater than 14, column c in the table would be used. The value in column c that corresponds to a Z score of 1.33 is 0.0918, so 9.18% of the scores in the distribution are greater than 14, which means that almost 10% of the months had more complaints recorded than the current month. Actually, because the area between the mean and 14 was known, it was not necessary to complete the second procedure. Because it was already determined that the percentage of scores between the mean and 14 was 40.82%, and because the area under the normal curve is always 100% and the area under half of the normal curve is 50%, the 40.08 could have been subtracted from 50 to arrive at the score of 9.18%—the area beyond that value.

These procedures work whether using scores above the mean or scores below the mean. As stated previously, 21 standard deviation is the same as 1 standard deviation; thus, it would be the same percentage if we were looking for the area under the normal curve representing less than 10. But what if you wanted to know how many scores were higher than one score and lower than another: for example, how many people

158 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 158 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

scored higher than 90 and lower than 62 on a test. This case would require determining the percentage greater than the score of 90 and the percentage less than the score of 62. The final pieces of information needed here are that the mean of the distribution is 76 and the standard deviation is 10.5. The areas beyond these two scores can be deter- mined simply by adding together the percentages in Table B-1 in Appendix B, “Statis- tical Tables.” The Z score calculations for these two values would be as follows:

Z 5 X 2 X

s Z 5

X 2 X s

5 90 2 76

10.5 5

62 2 76

10.5

5 2

1.5 5

1.5

5 1.33 5 21.33

The results of these calculations would be taken to Table B-1 in Appendix B, “Sta- tistical Tables.” Finding 1.33 in column a and wanting to determine the area under the normal curve beyond this value, we would look in column c. The value found in column c is 0.0918. This means that 9.18% of the scores are higher than 90. Since 62 has the same value, but negative, it also has a value in column c of 0.0918, so 9.18% of the scores are lower than 62. Adding these two values together would establish what percentage of people scored greater than 90 and less than 62:

0.0918 + 0.0918 = 0.1836

So 18.36% of the people in the class scored higher than 90 (making an A) or lower than 62 (making an F). It is also possible to determine the percentage of scores that fall between two scores. In the example we have been using, the percentage of scores that fall between 90 and 62 can be determined by adding together the areas between them and the mean. Looking at column b in Table B-1 in Appendix B, “Statistical Tables,” the percentage of scores between the mean of 76 and a score of 90 (the Z score of 1.33) is 40.82%. Since the score of 62 has the same Z score value, the percentage of scores between 62 and 76 is also 40.82%. Adding those percentages together produces a percentage of 81.64% as shown below:

0.4082 + 0.4082 = 0.8164

So 81.64% of the class scored a B, C, or D.

1596-6 Design of the Normal Curve

16304_CH06_Walker.indd 159 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Points to Remember about the Normal Curve

Two Z scores that will become important later are 1.96 and 2.58. These correspond to 95% and 99% of the area under the normal curve, respectively. These are important because researchers often want to know where 95% or 99% of the values in the distri- bution fall, or they may want to compare two values and see if they fall within 95% or 99% of the values in the distribution.

Another point you should remember is that the normal curve is a theoretical ideal. Normal distributions do not really exist, and the best we can do is to gather data that is close to a normal curve but not exact. The conclusions drawn when using the theory of the normal curve, then, will only be estimates, not hard facts.

Finally, you should realize that not all data sets even come close to a normal curve. In the discussion on measures of central tendency (see Chapter 4), some distri- butions are multimodal and almost jagged looking. There are many distributions that are so skewed they are J shaped and do not conform at all to a normal distribution. There are even a few distributions that are absolutely flat. Each of these distributions requires special treatment, which is discussed in the chapters on bivariate and multi- variate analysis.

7 Conclusion6- In this chapter we completed the description of one variable at a time from a distribution (univariate descriptive analysis) and you have learned how to describe a variable such that it can be relayed to another person accurately, and in a form that allows the person to get a mental image of the distribution. Describing single variables can take many forms: frequency distributions and graphs, measures of central tendency, measures of dispersion, and the form of the distribution. These all have the goal of describing the attributes of the distribution and determining if a variable is suitable for further analyses.

In the next chapters, we put together two variables—bivariate descriptive analy- ses—and describe what relationship might exist between them. The success of those analyses depends on the successful univariate description of data.

8 Key Terms6- form number of modes kurtosis platykurtic leptokurtic positively skewed mesokurtic skew moments symmetry negatively skewed Z score normal curve

160 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 160 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

9 Summary of Equations6- Moments of a Distribution

S1X 2 X2 i N

Z Score

Z 5 X 2 X

10 Exercises6- 1. Use the following frequency tables (from the gang database):

a. To discuss the number of modes. b. To determine the skew and kurtosis and discuss whether the distribution

is positively skewed or negatively skewed and whether it is leptokurtic, mesokurtic, or platykurtic.

HOME: What type of house do you live in?

Value Label Value Frequency Percent Valid

Percent Cumulative

Percent

House 1 280 81.6 82.4 82.4

Duplex 2 3 .9 .9 83.2

Trailer 3 34 9.9 10.0 93.2

Apartment 4 21 6.1 6.2 99.4

Other 5 2 .6 .6 100.0

Missing 3 .9

Total 343 100.0 100.0

N Valid 340

Missing 3

Mean 1.41

Std. Error of Mean .051

Median 1

Mode 1

Std. Deviation 0.945

Variance .892

Skewness 2.001

Std. Error of Skewness .132

Kurtosis 2.613

Std. Error of Kurtosis .264

Range 5

1616-10 Exercises

16304_CH06_Walker.indd 161 7/12/12 10:02:12 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

ARREST: How many times have you been arrested?

Value Frequency Percent Valid

Percent Cumulative

Percent

0 243 70.8 86.2 86.2

1 23 6.7 8.2 94.3

2 10 2.9 3.5 97.9

3 3 .9 1.1 98.9

5 2 .6 .7 99.6

24 1 .3 .4 100.0

Missing 61 17.8

Total 343 100.0 100.0

N Valid 282

Missing 61

Mean .30

Std. Error of Mean .093

Median 0

Mode 0

Std. Deviation 1.567

Variance 2.455

Skewness 12.692

Std. Error of Skewness .145

Kurtosis 187.898

Std. Error of Kurtosis .289

Range 24

TENURE: How long have you lived at your current address (months)?

Value Frequency Percent Valid

Percent Cumulative

Percent

1 14 4.1 4.3 4.3

2 6 1.7 1.8 6.1

3 4 1.2 1.2 7.3

4 4 1.2 1.2 8.6

162 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 162 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

TENURE: How long have you lived at your current address (months)?

Value Frequency Percent Valid

Percent Cumulative

Percent

5 6 1.7 1.8 10.4

6 6 1.7 1.8 12.2

7 1 .3 .3 12.5

8 3 .9 .9 13.5

9 2 .6 .6 14.1

10 1 .3 .3 14.4

11 1 .3 .3 14.7

12 11 3.2 3.4 18.0

14 1 .3 .3 18.3

18 5 1.5 1.5 19.9

21 1 .3 .3 20.2

24 30 8.7 9.2 29.4

30 1 .3 .3 29.7

31 1 .3 .3 30.0

32 1 .3 .3 30.3

36 22 6.4 6.7 37.0

42 1 .3 .3 37.3

48 12 3.5 3.7 41.0

60 24 7.0 7.3 48.3

72 14 4.1 4.3 52.6

76 1 .3 .3 52.9

84 8 2.3 2.4 55.4

96 18 5.2 5.5 60.9

108 4 1.2 1.2 62.1

120 9 2.6 2.8 64.8

132 11 3.2 3.4 68.2

144 21 6.1 6.4 74.6

156 13 3.8 4.0 78.6

168 11 3.2 3.4 82.0

1636-10 Exercises

16304_CH06_Walker.indd 163 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

TENURE: How long have you lived at your current address (months)?

Value Frequency Percent Valid

Percent Cumulative

Percent

170 5 1.5 1.5 83.5

180 7 2.0 2.1 85.6

182 2 .6 .6 86.2

186 1 .3 .3 86.5

192 14 4.1 4.3 90.8

198 1 .3 .3 91.1

204 24 7.0 7.3 98.5

216 3 .9 .9 99.4

240 2 .6 .6 100.0

Missing 16 4.7

Total 343 100.0 100.0

N Valid 327

Missing 16

Mean 88.77

Std. Error of Mean 3.880

Median 72

Mode 24

Std. Deviation 70.164

Variance 4923.055

Skewness .365

Std. Error of Skewness .135

Kurtosis 21.284

Std. Error of Kurtosis .269

Range 239

164 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 164 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

SIBS: How many brothers and sisters do you have?

Value Frequency Percent Valid Percent

Cumulative Percent

0 39 11.4 11.5 11.5

1 137 39.9 40.5 52.1

2 79 23.0 23.4 75.4

3 39 11.4 11.5 87.0

4 17 5.0 5.0 92.0

5 13 3.8 3.8 95.9

6 6 1.7 1.8 97.6

7 4 1.2 1.2 98.8

9 1 .3 .3 99.1

10 1 .3 .3 99.4

12 1 .3 .3 99.7

15 1 .3 .3 100.0

Missing 5 1.5

Total 343 100.0

N Valid 338

Missing 5

Mean 1.94

Std. Error of Mean .098

Median 1

Mode 1

Std. Deviation 1.801

Variance 3.245

Skewness 2.664

Std. Error of Skewness .133

Kurtosis 12.027

Std. Error of Kurtosis .265

Range 15

1656-10 Exercises

16304_CH06_Walker.indd 165 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

2. For a distribution with a mean of 50 and a standard deviation of 10: a. Calculate the Z score for a score of 40. b. Determine the area under the normal curve between the mean and this value. c. Determine the area under the normal curve beyond this value. d. Calculate a Z score for a score of 65. e. Determine the area under the normal curve between the mean and this value. f. Determine the area under the normal curve beyond this value. g. Determine the area under the normal curve between 40 and 60. h. Determine the area under the normal curve outside 40 and 60.

3. Say a parole board has a policy that it will only release prisoners who meet a minimum amount of time served, have a minimum number of good-time points, and have made an acceptable score on tests in their drug awareness class. If the mean of this distribution is 90, the minimum acceptable score on these criteria is 70, and the standard deviation of scores is 15: a. Calculate the Z score for a score of 90. b. Calculate the Z score for a score of 50. c. Determine the area under the normal curve between the mean and the val-

ues in parts a and b. d. Determine the area under the normal curve beyond these values. e. Calculate a Z score for a score of 70. f. Determine the area under the normal curve between the mean and this val-

ue. g. Determine the area under the normal curve beyond this value.

11 References6- MacGillivray, H. L. (1981). The mean, median, mode inequality and skew for a class

of densities. Australian Journal of Statistics, 23, 247.

12 For Further Reading6- Pearson, K. (1894). On the dissection of asymmetrical frequency-curves: General

theory. Philosophical Transactions of the Royal Society of London (Series A, Vol. 185). London: Cambridge University Press.

Pearson, K. (1895). Classification of asymmetrical frequency curves in general: Types actually occurring. Philosophical Transactions of the Royal Society of London (Series A, Vol. 186). London: Cambridge University Press.

166 Chapter 6 n The Form of a Distribution

16304_CH06_Walker.indd 166 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

13 Note6- 1. SPSS uses the convention of +1 to 21 as the measure of acceptable skew.

This is based on a particular formula that “standardizes” skew and kurtosis scores. There is a similar (and popular) formula that calculates skew and kur- tosis where the acceptable range is +3 to 23. Programs such as SAS use this formula. It is important, then, that you understand which formula is being used (what the acceptable range is) for these values before making a judgment about them.

Criminal Justice on the Web Visit http://criminaljustice.jbpub.com/Stats4e to make full use of today’s teaching and tech- nology! Our interactive Companion Website has been designed to specifically complement Statistics in Criminology and Criminal Justice: Analysis and Interpretation, 4th Edition. The resources available include a Glossary, Flashcards, Crossword Puzzles, Practice Quizzes, Web links, and Student Data Sets. Test yourself today!

1676-13 Note

16304_CH06_Walker.indd 167 7/12/12 10:02:13 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Table B-1 Area Under the Normal Curve

Appendix B

Statistical Tables

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

0.00 0.0000 0.5000 0.25 0.0987 0.4013 0.50 0.1915 0.3085 0.75 0.2734 0.2266 0.01 0.0040 0.4960 0.26 0.1026 0.3974 0.51 0.1950 0.3050 0.76 0.2764 0.2236 0.02 0.0080 0.4920 0.27 0.1064 0.3936 0.52 0.1985 0.3015 0.77 0.2794 0.2206 0.03 0.0120 0.4880 0.28 0.1103 0.3897 0.53 0.2019 0.2981 0.78 0.2823 0.2177 0.04 0.0160 0.4840 0.29 0.1141 0.3859 0.54 0.2054 0.2946 0.79 0.2852 0.2148 0.05 0.0199 0.4801 0.30 0.1179 0.3821 0.55 0.2088 0.2912 0.80 0.2881 0.2119 0.06 0.0239 0.4761 0.31 0.1217 0.3783 0.56 0.2123 0.2877 0.81 0.2910 0.2090 0.07 0.0279 0.4721 0.32 0.1255 0.3745 0.57 0.2157 0.2843 0.82 0.2939 0.2061 0.08 0.0319 0.4681 0.33 0.1293 0.3707 0.58 0.2190 0.2810 0.83 0.2967 0.2033 0.09 0.0359 0.4641 0.34 0.1331 0.3669 0.59 0.2224 0.2776 0.84 0.2995 0.2005 0.1 0.0398 0.4602 0.35 0.1368 0.3632 0.60 0.2257 0.2743 0.85 0.3023 0.1977 0.11 0.0438 0.4562 0.36 0.1406 0.3594 0.61 0.2291 0.2709 0.86 0.3051 0.1949 0.12 0.0478 0.4522 0.37 0.1443 0.3557 0.62 0.2324 0.2676 0.87 0.3078 0.1922 0.13 0.0517 0.4483 0.38 0.1480 0.3520 0.63 0.2357 0.2643 0.88 0.3106 0.1894 0.14 0.0557 0.4443 0.39 0.1517 0.3483 0.64 0.2389 0.2611 0.89 0.3133 0.1867 0.15 0.0596 0.4404 0.40 0.1554 0.3446 0.65 0.2422 0.2578 0.90 0.3159 0.1841 0.16 0.0636 0.4364 0.41 0.1591 0.3409 0.66 0.2454 0.2546 0.91 0.3186 0.1814 0.17 0.0675 0.4325 0.42 0.1628 0.3372 0.67 0.2486 0.2514 0.92 0.3212 0.1788 0.18 0.0714 0.4286 0.43 0.1664 0.3336 0.68 0.2517 0.2483 0.93 0.3238 0.1762 0.19 0.0753 0.4247 0.44 0.1700 0.3300 0.69 0.2549 0.2451 0.94 0.3264 0.1736 0.20 0.0793 0.4207 0.45 0.1736 0.3264 0.70 0.2580 0.2420 0.95 0.3289 0.1711 0.21 0.0832 0.4168 0.46 0.1772 0.3228 0.71 0.2611 0.2389 0.96 0.3315 0.1685 0.22 0.0871 0.4129 0.47 0.1808 0.3192 0.72 0.2642 0.2358 0.97 0.3340 0.1660 0.23 0.0910 0.4090 0.48 0.1844 0.3156 0.73 0.2673 0.2327 0.98 0.3365 0.1635 0.24 0.0948 0.4052 0.49 0.1879 0.3121 0.74 0.2704 0.2296 0.99 0.3389 0.1611

(continues)

501

16304_APPB_Walker.indd 501 7/7/12 9:59:19 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Table B-1 Area Under the Normal Curve (continued) a

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

1.00 0.3413 0.1587 1.45 0.4265 0.0735 1.90 0.4713 0.0287 2.35 0.4906 0.0094 1.01 0.3438 0.1562 1.46 0.4279 0.0721 1.91 0.4719 0.0281 2.36 0.4909 0.0091 1.02 0.3461 0.1539 1.47 0.4292 0.0708 1.92 0.4726 0.0274 2.37 0.4911 0.0089 1.03 0.3485 0.1515 1.48 0.4306 0.0694 1.93 0.4732 0.0268 2.38 0.4913 0.0087 1.04 0.3508 0.1492 1.49 0.4319 0.0681 1.94 0.4738 0.0262 2.39 0.4916 0.0084 1.05 0.3531 0.1469 1.50 0.4332 0.0668 1.95 0.4744 0.0256 2.40 0.4918 0.0082 1.06 0.3554 0.1446 1.51 0.4345 0.0655 1.96 0.4750 0.0250 2.41 0.4920 0.0080 1.07 0.3577 0.1423 1.52 0.4357 0.0643 1.97 0.4756 0.0244 2.42 0.4922 0.0078 1.08 0.3599 0.1401 1.53 0.4370 0.0630 1.98 0.4761 0.0239 2.43 0.4925 0.0075 1.09 0.3621 0.1379 1.54 0.4382 0.0618 1.99 0.4767 0.0233 2.44 0.4927 0.0073 1.10 0.3643 0.1357 1.55 0.4394 0.0606 2.00 0.4772 0.0228 2.45 0.4929 0.0071 1.11 0.3665 0.1335 1.56 0.4406 0.0594 2.01 0.4778 0.0222 2.46 0.4931 0.0069 1.12 0.3686 0.1314 1.57 0.4418 0.0582 2.02 0.4783 0.0217 2.47 0.4932 0.0068 1.13 0.3708 0.1292 1.58 0.4429 0.0571 2.03 0.4788 0.0212 2.48 0.4934 0.0066 1.14 0.3729 0.1271 1.59 0.4441 0.0559 2.04 0.4793 0.0207 2.49 0.4936 0.0064 1.15 0.3749 0.1251 1.60 0.4452 0.0548 2.05 0.4798 0.0202 2.50 0.4938 0.0062 1.16 0.3770 0.1230 1.61 0.4463 0.0537 2.06 0.4803 0.0197 2.51 0.4940 0.0060 1.17 0.3790 0.1210 1.62 0.4474 0.0526 2.07 0.4808 0.0192 2.52 0.4941 0.0059 1.18 0.3810 0.1190 1.63 0.4484 0.0516 2.08 0.4812 0.0188 2.53 0.4943 0.0057 1.19 0.3830 0.1170 1.64 0.4495 0.0505 2.09 0.4817 0.0183 2.54 0.4945 0.0055 1.20 0.3849 0.1151 1.65 0.4505 0.0495 2.10 0.4821 0.0179 2.55 0.4946 0.0054 1.21 0.3869 0.1131 1.66 0.4515 0.0485 2.11 0.4826 0.0174 2.56 0.4948 0.0052 1.22 0.3888 0.1112 1.67 0.4525 0.0475 2.12 0.4830 0.0170 2.57 0.4949 0.0051 1.23 0.3907 0.1093 1.68 0.4535 0.0465 2.13 0.4834 0.0166 2.58 0.4951 0.0049 1.24 0.3925 0.1075 1.69 0.4545 0.0455 2.14 0.4838 0.0162 2.59 0.4952 0.0048 1.25 0.3944 0.1056 1.70 0.4554 0.0446 2.15 0.4842 0.0158 2.60 0.4953 0.0047 1.26 0.3962 0.1038 1.71 0.4564 0.0436 2.16 0.4846 0.0154 2.61 0.4955 0.0045 1.27 0.3980 0.1020 1.72 0.4573 0.0427 2.17 0.4850 0.0150 2.62 0.4956 0.0044 1.28 0.3997 0.1003 1.73 0.4582 0.0418 2.18 0.4854 0.0146 2.63 0.4957 0.0043 1.29 0.4015 0.0985 1.74 0.4591 0.0409 2.19 0.4857 0.0143 2.64 0.4959 0.0041 1.30 0.4032 0.0968 1.75 0.4599 0.0401 2.20 0.4861 0.0139 2.65 0.4960 0.0040 1.31 0.4049 0.0951 1.76 0.4608 0.0392 2.21 0.4864 0.0136 2.66 0.4961 0.0039 1.32 0.4066 0.0934 1.77 0.4616 0.0384 2.22 0.4868 0.0132 2.67 0.4962 0.0038 1.33 0.4082 0.0918 1.78 0.4625 0.0375 2.23 0.4871 0.0129 2.68 0.4963 0.0037 1.34 0.4099 0.0901 1.79 0.4633 0.0367 2.24 0.4875 0.0125 2.69 0.4964 0.0036 1.35 0.4115 0.0885 1.80 0.4641 0.0359 2.25 0.4878 0.0122 2.70 0.4965 0.0035 1.36 0.4131 0.0869 1.81 0.4649 0.0351 2.26 0.4881 0.0119 2.71 0.4966 0.0034 1.37 0.4147 0.0853 1.82 0.4656 0.0344 2.27 0.4884 0.0116 2.72 0.4967 0.0033 1.38 0.4162 0.0838 1.83 0.4664 0.0336 2.28 0.4887 0.0113 2.73 0.4968 0.0032 1.39 0.4177 0.0823 1.84 0.4671 0.0329 2.29 0.4890 0.0110 2.74 0.4969 0.0031 1.40 0.4192 0.0808 1.85 0.4678 0.0322 2.30 0.4893 0.0107 2.75 0.4970 0.0030 1.41 0.4207 0.0793 1.86 0.4686 0.0314 2.31 0.4896 0.0104 2.76 0.4971 0.0029 1.42 0.4222 0.0778 1.87 0.4693 0.0307 2.32 0.4898 0.0102 2.77 0.4972 0.0028 1.43 0.4236 0.0764 1.88 0.4699 0.0301 2.33 0.4901 0.0099 2.78 0.4973 0.0027 1.44 0.4251 0.0749 1.89 0.4706 0.0294 2.34 0.4904 0.0096 2.79 0.4974 0.0026

(continues)

502 Appendix B n Statistical Tables

16304_APPB_Walker.indd 502 7/7/12 9:59:19 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Table B-1 Area Under the Normal Curve (continued) a

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

b Area

between X – and Z

c Area

beyond Z

2.80 0.4974 0.0026 2.95 0.4984 0.0016 3.10 0.4990 0.0010 3.25 0.4994 0.0006 2.81 0.4975 0.0025 2.96 0.4985 0.0015 3.11 0.4991 0.0009 3.30 0.4995 0.0005 2.82 0.4976 0.0024 2.97 0.4985 0.0015 3.12 0.4991 0.0009 3.35 0.4996 0.0004 2.83 0.4977 0.0023 2.98 0.4986 0.0014 3.13 0.4991 0.0009 3.40 0.4997 0.0003 2.84 0.4977 0.0023 2.99 0.4986 0.0014 3.14 0.4992 0.0008 3.45 0.4997 0.0003 2.85 0.4978 0.0022 3.00 0.4987 0.0013 3.15 0.4992 0.0008 3.50 0.4998 0.0002 2.86 0.4979 0.0021 3.01 0.4987 0.0013 3.16 0.4992 0.0008 3.60 0.4998 0.0002 2.87 0.4979 0.0021 3.02 0.4987 0.0013 3.17 0.4992 0.0008 3.70 0.4999 0.0001 2.88 0.4980 0.0020 3.03 0.4988 0.0012 3.18 0.4993 0.0007 3.80 0.4999 0.0001 2.89 0.4981 0.0019 3.04 0.4988 0.0012 3.19 0.4993 0.0007 3.90 0.49995 0.00005 2.90 0.4981 0.0019 3.05 0.4989 0.0011 3.20 0.4993 0.0007 4.00 0.49997 0.00003 2.91 0.4982 0.0018 3.06 0.4989 0.0011 3.21 0.4993 0.0007 4.50 0.4999966 0.0000034 2.92 0.4982 0.0018 3.07 0.4989 0.0011 3.22 0.4994 0.0006 5.00 0.4999997 0.0000003 2.93 0.4983 0.0017 3.08 0.4990 0.0010 3.23 0.4994 0.0006 5.50 0.4999999 0.0000001 2.94 0.4984 0.0016 3.09 0.4990 0.0010 3.24 0.4994 0.0006

Table B-2 Values of Chi-Square

Probability (Top Row) and Significance (Bottom Row)

0.999 0.99 0.95 0.90 0.80 0.70

0.0001 0.01 0.05 0.10 0.20 0.30

1 10.827 6.635 3.841 2.706 1.642 1.074

2 13.815 9.210 5.991 4.605 3.219 2.408

3 16.268 11.345 7.815 6.251 4.624 3.665

4 18.465 13.277 9.488 7.779 5.989 4.878

5 20.517 15.086 11.070 9.236 7.289 6.064

6 22.457 16.812 12.592 10.645 8.558 7.231

7 24.322 18.475 14.067 12.017 9.803 8.383

8 26.125 20.090 15.507 13.362 11.030 9.524

9 27.877 21.666 16.919 14.684 12.242 10.656

10 29.588 23.209 18.307 15.987 13.442 11.781

11 31.264 24.725 19.675 17.275 14.631 12.899

12 32.909 26.217 21.026 18.549 15.812 14.011

13 34.528 27.688 22.362 19.812 16.985 15.119

14 36.123 29.141 23.685 21.064 18.151 16.222

15 37.697 30.578 24.996 22.307 19.311 17.322

(continues)

Statistical Tables 503

16304_APPB_Walker.indd 503 7/7/12 9:59:20 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Table B-2 Values of Chi-Square (continued)

Probability (Top Row) and Significance (Bottom Row)

0.999 0.99 0.95 0.90 0.80 0.70

0.0001 0.01 0.05 0.10 0.20 0.30

16 39.252 32.000 26.296 23.542 20.465 18.418

17 40.790 33.409 27.587 24.769 21.615 19.511

18 42.312 34.805 28.869 25.989 22.760 20.601

19 43.820 36.191 30.144 27.204 23.900 21.689

20 45.315 37.566 31.410 28.412 25.038 22.775

21 46.797 38.932 32.671 29.615 26.171 23.858

22 48.268 40.289 33.924 30.813 27.301 24.939

23 49.728 41.638 35.172 32.007 28.429 26.018

24 51.179 42.980 36.415 33.196 29.553 27.096

25 52.620 44.314 37.652 34.382 30.675 28.172

26 54.052 45.642 38.885 35.563 31.795 29.246

27 55.476 46.963 40.113 36.741 32.912 30.319

28 56.893 48.278 41.337 37.916 34.027 31.391

29 58.302 49.588 42.557 39.087 35.139 32.461

30 59.703 50.892 43.773 40.256 36.250 33.530

504 Appendix B n Statistical Tables

16304_APPB_Walker.indd 504 7/7/12 9:59:20 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

Table B-3 Student’s t Distribution

Level of Significance for One-Tailed Test

0.10 0.05 0.025 0.01 0.005 0.0005

Level of Significance for Two-Tailed Test

df 0.20 0.10 0.05 0.02 0.01 0.001

1 3.078 6.314 12.706 31.821 63.657 636.620

2 1.886 2.920 4.303 6.965 9.925 31.598

3 1.638 2.353 3.182 4.541 5.841 12.941

4 1.533 2.132 2.776 3.747 4.604 8.610

5 1.476 2.015 2.571 3.365 4.032 6.859

6 1.440 1.943 2.447 3.143 3.707 5.959

7 1.415 1.895 2.365 2.998 3.499 5.405

8 1.397 1.860 2.306 2.896 3.355 5.041

9 1.383 1.833 2.262 2.821 3.250 4.781

10 1.372 1.812 2.228 2.764 3.169 4.587

11 1.363 1.796 2.201 2.718 3.106 4.437

12 1.356 1.782 2.179 2.681 3.055 4.318

13 1.350 1.771 2.160 2.650 3.012 4.221

14 1.345 1.761 2.145 2.624 2.977 4.140

15 1.341 1.753 2.131 2.602 2.947 4.073

16 1.337 1.746 2.120 2.583 2.921 4.015

17 1.333 1.740 2.110 2.567 2.898 3.965

18 1.330 1.734 2.101 2.552 2.878 3.922

19 1.328 1.729 2.093 2.539 2.861 3.883

20 1.325 1.725 2.086 2.528 2.845 3.850

21 1.323 1.721 2.080 2.518 2.831 3.819

22 1.321 1.717 2.074 2.508 2.819 3.792

23 1.319 1.714 2.069 2.500 2.807 3.767

24 1.318 1.711 2.064 2.492 2.797 3.745

25 1.316 1.708 2.060 2.485 2.787 3.725

26 1.315 1.706 2.056 2.479 2.779 3.707

27 1.314 1.703 2.052 2.473 2.771 3.690

28 1.313 1.701 2.048 2.467 2.763 3.674

29 1.311 1.699 2.045 2.462 2.756 3.659

30 1.310 1.697 2.042 2.457 2.750 3.646

40 1.303 1.684 2.021 2.423 2.704 3.551

60 1.296 1.671 2.000 2.390 2.660 3.460

120 1.289 1.658 1.980 2.358 2.617 3.373

∞ 1.282 1.645 1.960 2.326 2.576 3.291

(continues)

Statistical Tables 505

16304_APPB_Walker.indd 505 7/7/12 9:59:20 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U

T ab

le B

-4

D is

tr ib

ut io

n of

f ;

p =

.0 5

d f1

= 1

2 3

4 5

6 7

8 9

10 12

15 20

24 30

40 60

12 0

∞ d

f2 =

1 16

1. 44

76 19

9. 5

21 5.

70 73

22 4.

58 32

23 0.

16 19

23 3.

98 6

23 6.

76 84

23 8.

88 27

24 0.

54 33

24 1.

88 17

24 3.

90 6

24 5.

94 99

24 8.

01 31

24 9.

05 18

25 0.

09 51

25 1.

14 32

25 2.

19 57

25 3.

25 29

25 4.

31 44

2 18

.5 12

8 19

19 .1

64 3

19 .2

46 8

19 .2

96 4

19 .3

29 5

19 .3

53 2

19 .3

71 19

.3 84

8 19

.3 95

9 19

.4 12

5 19

.4 29

1 19

.4 45

8 19

.4 54

1 19

.4 62

4 19

.4 70

7 19

.4 79

1 19

.4 87

4 19

.4 95

7 3

10 .1

28 9.

55 21

9. 27

66 9.

11 72

9. 01

35 8.

94 06

8. 88

67 8.

84 52

8. 81

23 8.

78 55

8. 74

46 8.

70 29

8. 66

02 8.

63 85

8. 61

66 8.

59 44

8. 57

2 8.

54 94

8. 52

64 4

7. 70

86 6.

94 43

6. 59

14 6.

38 82

6. 25

61 6.

16 31

6. 09

42 6.

04 1

5. 99

88 5.

96 44

5. 91

17 5.

85 78

5. 80

25 5.

77 44

5. 74

59 5.

71 7

5. 68

77 5.

65 81

5. 62

81 5

6. 60

79 5.

78 61

5. 40

95 5.

19 22

5. 05

03 4.

95 03

4. 87

59 4.

81 83

4. 77

25 4.

73 51

4. 67

77 4.

61 88

4. 55

81 4.

52 72

4. 49

57 4.

46 38

4. 43

14 4.

39 85

4. 36

6 5.

98 74

5. 14

33 4.

75 71

4. 53

37 4.

38 74

4. 28

39 4.

20 67

4. 14

68 4.

09 9

4. 06

3. 99

99 3.

93 81

3. 87

42 3.

84 15

3. 80

82 3.

77 43

3. 73

98 3.

70 47

3. 66

89 7

5. 59

14 4.

73 74

4. 34

68 4.

12 03

3. 97

15 3.

86 6

3. 78

7 3.

72 57

3. 67

67 3.

63 65

3. 57

47 3.

51 07

3. 44

45 3.

41 05

3. 37

58 3.

34 04

3. 30

43 3.

26 74

3. 22

98 8

5. 31

77 4.

45 9

4. 06

62 3.

83 79

3. 68

75 3.

58 06

3. 50

05 3.

43 81

3. 38

81 3.

34 72

3. 28

39 3.

21 84

3. 15

03 3.

11 52

3. 07

94 3.

04 28

3. 00

53 2.

96 69

2. 92

76 9

5. 11

74 4.

25 65

3. 86

25 3.

63 31

3. 48

17 3.

37 38

3. 29

27 3.

22 96

3. 17

89 3.

13 73

3. 07

29 3.

00 61

2. 93

65 2.

90 05

2. 86

37 2.

82 59

2. 78

72 2.

74 75

2. 70

67 10

4. 96

46 4.

10 28

3. 70

83 3.

47 8

3. 32

58 3.

21 72

3. 13

55 3.

07 17

3. 02

04 2.

97 82

2. 91

3 2.

84 5

2. 77

4 2.

73 72

2. 69

96 2.

66 09

2. 62

11 2.

58 01

2. 53

11 4.

84 43

3. 98

23 3.

58 74

3. 35

67 3.

20 39

3. 09

46 3.

01 23

2. 94

8 2.

89 62

2. 85

36 2.

78 76

2. 71

86 2.

64 64

2. 60

9 2.

57 05

2. 53

09 2.

49 01

2. 44

8 2.

40 45

12 4.

74 72

3. 88

53 3.

49 03

3. 25

92 3.

10 59

2. 99

61 2.

91 34

2. 84

86 2.

79 64

2. 75

34 2.

68 66

2. 61

69 2.

54 36

2. 50

55 2.

46 63

2. 42

59 2.

38 42

2. 34

1 2.

29 62

13 4.

66 72

3. 80

56 3.

41 05

3. 17

91 3.

02 54

2. 91

53 2.

83 21

2. 76

69 2.

71 44

2. 67

1 2.

60 37

2. 53

31 2.

45 89

2. 42

02 2.

38 03

2. 33

92 2.

29 66

2. 25

24 2.

20 64

14 4.

60 01

3. 73

89 3.

34 39

3. 11

22 2.

95 82

2. 84

77 2.

76 42

2. 69

87 2.

64 58

2. 60

22 2.

53 42

2. 46

3 2.

38 79

2. 34

87 2.

30 82

2. 26

64 2.

22 29

2. 17

78 2.

13 07

15 4.

54 31

3. 68

23 3.

28 74

3. 05

56 2.

90 13

2. 79

05 2.

70 66

2. 64

08 2.

58 76

2. 54

37 2.

47 53

2. 40

34 2.

32 75

2. 28

78 2.

24 68

2. 20

43 2.

16 01

2. 11

41 2.

06 58

16 4.

49 4

3. 63

37 3.

23 89

3. 00

69 2.

85 24

2. 74

13 2.

65 72

2. 59

11 2.

53 77

2. 49

35 2.

42 47

2. 35

22 2.

27 56

2. 23

54 2.

19 38

2. 15

07 2.

10 58

2. 05

89 2.

00 96

17 4.

45 13

3. 59

15 3.

19 68

2. 96

47 2.

81 2.

69 87

2. 61

43 2.

54 8

2. 49

43 2.

44 99

2. 38

07 2.

30 77

2. 23

04 2.

18 98

2. 14

77 2.

10 4

2. 05

84 2.

01 07

1. 96

04 18

4. 41

39 3.

55 46

3. 15

99 2.

92 77

2. 77

29 2.

66 13

2. 57

67 2.

51 02

2. 45

63 2.

41 17

2. 34

21 2.

26 86

2. 19

06 2.

14 97

2. 10

71 2.

06 29

2. 01

66 1.

96 81

1. 91

68 19

4. 38

07 3.

52 19

3. 12

74 2.

89 51

2. 74

01 2.

62 83

2. 54

35 2.

47 68

2. 42

27 2.

37 79

2. 30

8 2.

23 41

2. 15

55 2.

11 41

2. 07

12 2.

02 64

1. 97

95 1.

93 02

1. 87

8 20

4. 35

12 3.

49 28

3. 09

84 2.

86 61

2. 71

09 2.

59 9

2. 51

4 2.

44 71

2. 39

28 2.

34 79

2. 27

76 2.

20 33

2. 12

42 2.

08 25

2. 03

91 1.

99 38

1. 94

64 1.

89 63

1. 84

21 4.

32 48

3. 46

68 3.

07 25

2. 84

01 2.

68 48

2. 57

27 2.

48 76

2. 42

05 2.

36 6

2. 32

1 2.

25 04

2. 17

57 2.

09 6

2. 05

4 2.

01 02

1. 96

45 1.

91 65

1. 86

57 1.

81 17

22 4.

30 09

3. 44

34 3.

04 91

2. 81

67 2.

66 13

2. 54

91 2.

46 38

2. 39

65 2.

34 19

2. 29

67 2.

22 58

2. 15

08 2.

07 07

2. 02

83 1.

98 42

1. 93

8 1.

88 94

1. 83

8 1.

78 31

23 4.

27 93

3. 42

21 3.

02 8

2. 79

55 2.

64 2.

52 77

2. 44

22 2.

37 48

2. 32

01 2.

27 47

2. 20

36 2.

12 82

2. 04

76 2.

00 5

1. 96

05 1.

91 39

1. 86

48 1.

81 28

1. 75

7 24

4. 25

97 3.

40 28

3. 00

88 2.

77 63

2. 62

07 2.

50 82

2. 42

26 2.

35 51

2. 30

02 2.

25 47

2. 18

34 2.

10 77

2. 02

67 1.

98 38

1. 93

9 1.

89 2

1. 84

24 1.

78 96

1. 73

3 25

4. 24

17 3.

38 52

2. 99

12 2.

75 87

2. 60

3 2.

49 04

2. 40

47 2.

33 71

2. 28

21 2.

23 65

2. 16

49 2.

08 89

2. 00

75 1.

96 43

1. 91

92 1.

87 18

1. 82

17 1.

76 84

1. 71

26 4.

22 52

3. 36

9 2.

97 52

2. 74

26 2.

58 68

2. 47

41 2.

38 83

2. 32

05 2.

26 55

2. 21

97 2.

14 79

2. 07

16 1.

98 98

1. 94

64 1.

90 1

1. 85

33 1.

80 27

1. 74

88 1.

69 06

27 4.

21 3.

35 41

2. 96

04 2.

72 78

2. 57

19 2.

45 91

2. 37

32 2.

30 53

2. 25

01 2.

20 43

2. 13

23 2.

05 58

1. 97

36 1.

92 99

1. 88

42 1.

83 61

1. 78

51 1.

73 06

1. 67

17 28

4. 19

6 3.

34 04

2. 94

67 2.

71 41

2. 55

81 2.

44 53

2. 35

93 2.

29 13

2. 23

6 2.

19 2.

11 79

2. 04

11 1.

95 86

1. 91

47 1.

86 87

1. 82

03 1.

76 89

1. 71

38 1.

65 41

29 4.

18 3

3. 32

77 2.

93 4

2. 70

14 2.

54 54

2. 43

24 2.

34 63

2. 27

83 2.

22 29

2. 17

68 2.

10 45

2. 02

75 1.

94 46

1. 90

05 1.

85 43

1. 80

55 1.

75 37

1. 69

81 1.

63 76

30 4.

17 09

3. 31

58 2.

92 23

2. 68

96 2.

53 36

2. 42

05 2.

33 43

2. 26

62 2.

21 07

2. 16

46 2.

09 21

2. 01

48 1.

93 17

1. 88

74 1.

84 09

1. 79

18 1.

73 96

1. 68

35 1.

62 23

40 4.

08 47

3. 23

17 2.

83 87

2. 60

6 2.

44 95

2. 33

59 2.

24 9

2. 18

02 2.

12 4

2. 07

72 2.

00 35

1. 92

45 1.

83 89

1. 79

29 1.

74 44

1. 69

28 1.

63 73

1. 57

66 1.

50 89

60 4.

00 12

3. 15

04 2.

75 81

2. 52

52 2.

36 83

2. 25

41 2.

16 65

2. 09

7 2.

04 01

1. 99

26 1.

91 74

1. 83

64 1.

74 8

1. 70

01 1.

64 91

1. 59

43 1.

53 43

1. 46

73 1.

38 93

12 0

3. 92

01 3.

07 18

2. 68

02 2.

44 72

2. 28

99 2.

17 5

2. 08

68 2.

01 64

1. 95

88 1.

91 05

1. 83

37 1.

75 05

1. 65

87 1.

60 84

1. 55

43 1.

49 52

1. 42

9 1.

35 19

1. 25

39 ∞

3. 84

15 2.

99 57

2. 60

49 2.

37 19

2. 21

41 2.

09 86

2. 00

96 1.

93 84

1. 87

99 1.

83 07

1. 75

22 1.

66 64

1. 57

05 1.

51 73

1. 45

91 1.

39 4

1. 31

8 1.

22 14

506 Appendix B n Statistical Tables

16304_APPB_Walker.indd 506 7/7/12 9:59:20 AM

R O D D Y , A N T H O N Y I S A A C 3 7 2 7 B U