Easterly-Smith et al (2008:252) 4-step hypothesis testing will be a guide my analysis of Grandson’s et al (2003) hypothesis as it takes Into account the general steps to hypothesis testing. ; Step 1 (Define research hypothesis): Grandson’s et al (2003) defined three hypotheses for analysis as stated above, which meets the criteria. ; Step 2 (Define Null Hypothesis): There was no Null Hypothesis stated to be tested. Although the null hypothesis can be assumed It is better for it to be stated for testing. Barman M.
Even (2006:149) notes, “an objective of statistical testing Is examine whether the null hypothesis can be rejected”. ; Step 3: The authors derived a summary measure by analyzing the mean, standard deviation and standardized sidedness (Grandson’s et al 2003). From the analysis the HI was rejected as a result of differences found between a non-clinicians number (1 1 out of ; Step 4: the authors also used Chi-Square testing as one of their 38 questions). Reference distribution and test statistic to measure relationship between the two variables, which was appropriate in these instances.
From the above I can argue that though the authors sampling and data used appropriate measurement steps with conclusions drawn from the results,
Need essay sample on "Management Research"? We will write a custom essay sample specifically for you for only $ 13.90/page
Grandson’s et al (2003) then posit that by using the sampling design suggested, the sample became biased as experienced Internet users became more prominent in the data collection and their expertise was noted during analysis. This showed in their I OFF were question that expert Internet users found in the web survey sampling were confident to answer. What might you have suggested to them to improve this argument? I would have suggested that the researcher use a “simplification through selecting a sub-sample” to deal with interrelated factors, where Easterly-Smith et al (2008:272) notes that samples should be equal in both groups.
That is sample a population of expert Internet users or intermediary users, making sure that the “potentially confounding factors” are equal, in this case will be the expertise of the sample. If they state their null hypothesis and their argument would have been to accept or reject the null hypothesis during hypothesis. Though this might not reflect on their conclusions of sample bias and mode of administration, it would have made the research more focused on its intentions to accept or reject the Null or Alternative hypotheses.
What other criticisms can you offer? Given that the authors had used the sampling design that was not reflective of their results and seemed like without a Null hypothesis abandoned their research to conclude on the issue of sample bias, the researchers could have modified their impaling design after they realized that their sampling was bias and gathered data to reflect an expert pool of Internet users for the paper survey to match those that were collected for web survey to conclude their analyses.
To combat low response rate especially for students survey, the authors should have given an incentive for participation, sent reminders and send the survey as email as opposed to a URL as non experts are weary of virus and spam mails. I believe the length of the survey and some of the questions were too long and not engaging enough for a survey targeted at students. Hankerer and Crayon (2009:360) suggest that pre-notification, length of survey and constant reminders through phone calls and emails should be sent to help boost response rate.