I have confusion between the two statistical tests that are to be applied to test the hypothesis. The hypothesis to be tested is (H0) There is no significant impact of e-banking on bank's performance. Impact of e-banking on bank performance is measured using a five point likert scale. Questions were asked to bank employees regarding the same. I was supposed to apply one sample t -test but the sample data is not normal. So, should I apply a non-parametric test since the data violates the parametric test assumption ( in this case one sample Wilcoxon signed test). The sample size of the study is 384.
Shikha Agrawal wrote: "The hypothesis to be tested is (H0) There is no significant impact of e-banking on bank's performance."
That statement suggests that several banks are the units of observation, and that for each bank, you have a measure of e-banking, another measure of performance, and that you wish to examine the association between those two variables. At least one of those variables appears to be measured using a 5-point ordinal item. Perhaps both are. Please clarify. One possibility might be the ordinal Chi-square test that goes by various names. You can read about it here:
- https://www.uvm.edu/~statdhtx/StatPages/More_Stuff/OrdinalChisq/OrdinalChiSq.html
I did not understand what you were getting at when you added this: "Questions were asked to bank employees regarding the same." Who provided the measurements described above? Are you asking if the nature of the association between variables depends on who takes the "measurements"?
HTH.
Dear Shikha Agrawal ,
Kindly explain more accurately how "Impact of e-banking on bank performance" is measured. Is e-banking measured by a single Likert question or a group of questions? Likewise for bank performance? Or are you only measuring "Impact of e-banking on bank performance" using one or more questions? The answer to your question depends upon this.
Generally, the impact of A upon B is measured with a correlation. If a scatter plot of A against B is not "cigar shaped" you should use Spearman rank correlation. Please see my study guides.
In my opinion, based on the central limit theorem, you can justify that your sample is normal and use parametric tests.
Respected Sir
I would like to clear my question. I have measure the impact of e-banking on bank performance by asking group of questions to bank employees. The questions were based framed on five point likert scale. Questions like E-banking reduces cost of transaction, e-banking reduces complexity in processing , e-banking reduces human error etc. I have summated the likert score of the questions found reliable and valid by CFA test on AMOS. Now please suggest me which test I should apply.
After reviewing the previous literature, I think one sample t test taking test value as 3 (i.e. neutral) and in case of non normality of data one sample wilcoxon rank test will be suitable. But I am not confident and still confused . Please suggest me the correct test ?
Thank You
This appears to be a homework question. Has the policy changed on that?
D. Booth
Shikha Agrawal
You summarized bank performance in a single variable evaluated only by bank employees. It is then about the employees' perception of the bank's performance (different from what the performance itself may be). On the one hand, it is curious that everything is summarized in a single global score, without giving information structured in factors, for example, of an administrative nature, customer relations, market positioning, etc. On the other hand, we do not know if the survey was applied in several banks or in only one, or even in a single branch.
Assuming that the above is resolved, the hypothesis to be tested establishes that e-banking improves the overall perception of employees about the bank's performance. The scores that favor this are those above 3, which you mention as the perception of neutrality. Thus, your alternative hypothesis is
Probability (Global perception> 3)> Probability (Global perception 3)
Thank You Sir
Sir, there are parameters which are as follow :
customer satisfaction, bank service quality, productivity, service processing time and operational efficiency
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size