I would like to, first of all, congratulate scholars at the Department of Political Science, University of Ghana, for their efforts, time and resources in the design and implementation of this very important survey.
This has indeed ignited some discussions about the performance of Ghanaian parliamentarians in the discharge of their duties and a wake-up call to some of them.
However, other scholars would have preferred an objective measurement of performance, yet this exercise is also relevant in complementing a purely quantitative approach and when such even does not exist, a perception-driven survey, like this one, becomes very necessary. Therefore, going forward, a debate could ensue and consensus be built on how best parliamentarians and government appointees could be assessed, somewhat, objectively.
Notwithstanding the commendations to the scientists for a generally good work done, there are very essential issues of concern that would bother on credibility and verifiability of findings if our convictions are right. The issues to be pointed out are very crucial to both users and producers of scientific knowledge. Once more, we are not casting doubts on the truthfulness of the findings but only to indicate few issues which have the potency to damage the credibility and verifiability of findings, perhaps, due to the manner they have been reported. Some of these issues mildly start from the sampling approach up to the conclusions drawn on the findings.
To begin with, the report stated that a mixed method was adopted where qualitative data was used to validate findings by the quantitative approach. However, this is evidently missing in the report. That is, the reader is not shown where qualitative data was used to validate a quantitative finding. This raises questions about the stated method. Also, the report slides indicated that, “Interviews are conducted at the appropriate electoral areas and constituencies to solicit key information to validate the quantitative data. In each Electoral area, 20 respondents were interviewed. Therefore, with 5 electoral areas per constituency and (100) were engaged”. Were all respondents interviewed? Again, at the methodology section where the report indicates that the five (5) electoral areas where selected using the hat simple random approach, the reader is not told how the individual respondents were particularly sampled. Could they have been sampled by simple random too? These are pretty petty omissions!
The most important challenge one would find is the manner in which percentages have been reported and without statistical tests. For almost all tables presented in the opinion poll document, it is difficult to tell whether they are row or column percentages and whether they were calculated based on the regional sample sizes or the countrywide sample. For instance, on pages 10 and 11 where some respondents indicated that MPs have fulfilled their campaign promises on a set of items while, on the same items of campaign promises, some respondents stated that they have not been fulfilled. Yet, adding the percentages of these two opposing views does not add up to 100%, likewise across the promised items. Were there missing observations which were not reported?
The lack of statistical test on the mere differences among categories of variables makes it difficult for anyone to conclude sharply about any identifiable differences in the numbers. So, the conclusion that, “But from our study, while a third of the long-term MPs (in their fourth-seventh) term still have more than 50% support of their constituents to seek election again in 2020, two thirds of the first term MPs do not have such support”, must be tested empirically. These are brief but have grave implications on the entire work done.
It is important to reiterate that the purpose of this scrutiny is not to castigate the entire research but to highlight few challenges that authors of the report could have paid attention to with keen interest. These issues raised in here relate to the credibility of finding and as academic community, we would do so much good to society if we would report credible and verifiable research findings. After all, that is all that science of research is about!
Source: Richard Kwabena Nkrumah | Innovations for Data Analytics and Policy |idaponline.org