WK 5 dis Data

These are results from a student’s DNP project. The student conducted completed Rapid Response Team (RRT) calls asking activating nurses to respond to seven questions about the RRT, indicating their level of satisfaction with the call. Each of the seven questions was measured on a 5-point scale where 1 indicated strong dissatisfaction with the RRT process and 5 equaled a very strong satisfaction with the way the RRT was conducted. Thus, the highest possible score on the RRT evaluation survey was 35 (5 x 7), and the lowest possible score was 7 (i.e., 7 x 1). Higher scores indicate higher satisfaction with the RRT response, and lower scores indicate lower satisfaction with the RRT response.

The data was initially compiled in Excel but then transferred to statistical software for analysis. Selected tests are provided below with explanations.

Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other

The information below is a frequency table with counts. There were 25 cases; of these, 14 were cardiac, three were respiratory, and seven were neurology in nature. There was one RRT call that was for another reason, and this case was removed from the analysis.

Case Processing Summary
	Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other	Cases
		Valid	Missing	Total
		N	Percent	N	Percent	N	Percent
RRTEvalSum	cardiac	14	100.0%	0	0.0%	14	100.0%
	respiratory	3	100.0%	0	0.0%	3	100.0%
	neurology	7	100.0%	0	0.0%	7	100.0%
	all other	1	100.0%	0	0.0%	1	100.0%

The student wanted to see if there was a statistical difference in the mean scores on the RRT evaluation survey. When the responses on the survey were summarized below. There are descriptive statistics showcasing measures of central tendency (the mean and median) as well as measures of dispersion like the standard deviation, variance, and range.

Descriptivesa
	Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other	Statistic	Std. Error
RRTEvalSum	cardiac	Mean	30.0714	1.52800
		95% Confidence Interval for Mean	Lower Bound	26.7704
			Upper Bound	33.3725
		5% Trimmed Mean	30.6349
		Median	31.0000
		Variance	32.687
		Std. Deviation	5.71724
		Minimum	15.00
		Maximum	35.00
		Range	20.00
		Interquartile Range	7.25
		Skewness	-1.651	.597
		Kurtosis	2.867	1.154
	respiratory	Mean	29.6667	1.76383
		95% Confidence Interval for Mean	Lower Bound	22.0775
			Upper Bound	37.2558
		5% Trimmed Mean	.
		Median	29.0000
		Variance	9.333
		Std. Deviation	3.05505
		Minimum	27.00
		Maximum	33.00
		Range	6.00
		Interquartile Range	.
		Skewness	.935	1.225
		Kurtosis	.	.
	neurology	Mean	28.2857	3.90795
		95% Confidence Interval for Mean	Lower Bound	18.7233
			Upper Bound	37.8481
		5% Trimmed Mean	29.0952
		Median	33.0000
		Variance	106.905
		Std. Deviation	10.33948
		Minimum	7.00
		Maximum	35.00
		Range	28.00
		Interquartile Range	12.00
		Skewness	-1.852	.794
		Kurtosis	3.284	1.587
a. RRTEvalSum is constant when Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other = all other. It has been omitted. Note that cardiac RRTs had the most positive satisfaction score (30.0714), with a median of 31. Respiratory had a mean score of 29.6667 with a median of 29. Neurology was last with a mean score of 28.2857 and a median of 33, indicating some distortion in the distribution of the data. This likely due to the very small sample size, which is distorted even more when segmented into three groups. Let’s see, nonetheless, if there was a statistically significant difference in the mean between these three groups by using the nonparametric equivalent of the one-way ANOVA, the Kruskal-Wallis.

Kruskal-Wallis Test

Ranks
	Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other	N	Mean Rank
RRTEvalSum	cardiac	14	12.86
	respiratory	3	9.67
	neurology	7	13.50
	all other	1	21.50
	Total	25

Test Statisticsa,b
	RRTEvalSum
Kruskal-Wallis H	2.059
df	3
Asymp. Sig.	.560
a. Kruskal-Wallis Test
b. Grouping Variable: Description of the Call 1= cardiac 2= respiratory 3= neurology 4= other

The statistical software discounts one case in the “all other” category and compares cardiac, respiratory, and neurology. The H statistic used in the statistical software was 2.059 with 3 degrees of freedom and an asymptotic significance level of .560. As this result is far greater than the accepted .05 level of significance, there is no point in going any further. The difference in the means is not statistically significant.

Degrees of freedom have to do with (essentially) an estimate of the sample size. In this case, all four groups are counted, yielding 3 degrees of freedom (1-#of groups).

Salkind and Frey’s 5th edition has a good explanation of effect size (the magnitude of the difference) on pages 243–246.

Read about the word asymptotic on page 183.

Now, because the sample size is so very small, the student took pains to explain that with a small between the groups, she would have needed a huge sample—in fact, over 250—in order to see whether this small difference was statistically significant or not.

Thus, this very small dataset exhibits a very likely case of a type II error when significance was not found, but it may, in fact, be there.