PROTEACH INFO

Rules for Forming Hypotheses A (alternative) hypothesis is a statement of what you believe based on deductive reasoning. The null hypothesis, which is the opposite of the hypothesis, is tested in hopes that it can be REJECTED, thereby implying the other hypothesis can be supported (NOTICE we do not say true, false or proven). In journal articles, if only one hypothesis is shown, it is usually the HYPOTHESIS. We are really interested in the hypothesis, but the rules of statistics dictate that we test the null hypothesis. You only test concepts that are measured by your Surveys (the FACTORS**) A survey is made up of questions. The questions will either measure a demographic (a label describing a person, thing). Examples would be gender, education, age, tenure, etc. OR They are questions that when put together (either averaged or summed) measure an abstract concept…we call this a scale or factor score (The individual portion of the U4 Group Assignment). When writing hypotheses, you do not compare on a single question, but rather a concept or factor/scale. If you measure a single question in a hypothesis for the project, you will get the whole question wrong. Each hypothesis must contain a comparison of one of the factors in your scale. You can compare two different factors or a factor plus a demographic (for example). Wording for ANOVAs & T-Tests: NULL: Males are the same as females with regard to ____________________. HYPO: Males are not the same as females with regard to ____________________.

Three Possible Statements of Hypotheses

HYPOTHESIS NULL HYPOTHESIS

LOWER TAIL Less than < Greater than or equal to >/=

UPPER TAIL Greater than > Less than or equal to </=

TWO TAIL Not equal to =/ Equal to =

NOTE: although in advanced statistical testing, an equality symbol may be found in either

the hypothesis or the null, it is often easier to have the equality sign in the NULL HYPOTHESIS. You may set it up either way, but the preferred manner (at this stage) is stated

in the table above.

WORDING FOR DECISION RULE….

These are not tests, but words to describe the Reject/Do Not Reject Status p-VALUE approach Given that the sig. (xx) is greater than the alpha (.xx), the NULL cannot be rejected therefore there is no support for the HYPO that (paste HYPO here) OR Given that the sig. (xx) is less than the alpha (.xx), the NULL is rejected therefore there is support for the HYPO that (paste HYPO here) Let us NOW look for the wording of the decision rule

EXAMPLE: A TOTALLY DIFFERENT SURVEY IS BEING USED…. Given that the sig. xx is less than the alpha of .05, the NULL hypothesis is rejected and therefore there is support for the HYPO that (insert HYPO) Given that the sig. xx is greater than the alpha of .05, the NULL hypothesis is not rejected and therefore there is no support for the HYPO that (insert HYPO) Don’t get fancy and start using words like the HYPO is accepted or is true…just stay with these simple phrases…and remember-there are no absolutes in this HYPO testing game…that is why we use the concept of SUPPORT for a hypothesis.

NULL: Males have the same level of Overall Job Satisfaction compared to Females. (M = F)

Males have a different level of Overall Job Satisfaction compared to Females. (M =/ F)

Looking at the mean Job Satisfaction scores for both genders shows that they are nearly equal, though the standard deviation for females is much larger, showing that the scores are less consistent for females.

Group Statistics

24 2.7361 .18768 .03831

83 2.7925 .67559 .07416

Gender

Male

Female

Overall Job Satisfaction

N Mean Std. Deviat ion Std. Error Mean

Given that the sig. .688 is greater than the alpha of .05, the NULL hypothesis is not rejected and therefore there is no support for the HYPO that Males have a different level of Overall Job Satisfaction compared to Females. (M =/ F).

We can also run an ANOVA to test this:

Independent Samples Test

37.249 .000 -.403 105 .688 -.0564 .13986 -.33371 .22092

-.676 105.0 .501 -.0564 .08347 -.22189 .10911

Equal variances

assumed

Equal variances

not assumed

Overall Job

Satisfaction

F Sig.

Levene's Test

for Equal ity of

Variances

t df

Sig.

(2-tail

ed)

Mean

Differen

ce

Std.

Error

Differ

ence Lower Upper

95% Confidence

Interval of the

Difference

t-test for Equality of Means

ANOVA

Overall Job Satisfaction

.059 1 .059 .163 .688

38.237 105 .364

38.296 106

Between

Groups

Within Groups

Total

Sum of Squares df Mean Square F Sig.

Given that the sig. .688 is greater than the alpha of .05, the NULL hypothesis is not rejected and therefore there is no support for the HYPO that Males have a different level of Overall Job Satisfaction compared to Females. (M =/ F).

MKTG420_U2IP.doc