Download figures to accopmany this chapter HERE

Figure 4.2 shows the frequencies for each of the quality of interactions within the three ordinal groups, and table 4.1 shows the summary statistics for the three groups. Of the total sample of 8,415 respondents, 53.6% were in the high quality of interaction group while 44.1% were in the medium group. A very small proportion of respondents, 2.3%, rated the overall quality of interactions at their institution as low. There were no significant differences in gender, with 52.2% of females rating their interactions as high and 54.4% of males doing so. Older age groups were more likely to rate their interactions higher, with 62.8% of those 30 years and older having a high rating while 51.3% of those aged 19/20 years doing so. Significant differences were present among different fields of study; 61.2% of those studying agriculture rated their interactions as high, and 48.5% of those studying business did so. Among institution types, students at Other Institutions (61.0%) were more likely to have a high rating of quality of interactions compared with students at Technological HE Institutions (55.7%) and Universities (49.3%).

We now turn to examining differences in the three outcome measures among the three quality of interaction groups. The first measure of student’s overall evaluation of their institution, with responses groups as either positive or negative evaluation.

Figure 4.3 shows that among students who reported high quality of interactions, 92.2% had a positive overall evaluation of their institution. Among those in the medium group, a significantly smaller proportion (69.8%) did so, while among students who rated the quality of interactions as low, 40.8% reported that they had a positive evaluation of their institution.

The second outcome measure considered responses to the question “if you could start over again, would you go to the same institution you are now attending?”. 92.6% of those in the high quality of interaction group said that they would do so. A significantly lower proportion in the medium group, 80.0%, said that they would, while 53.2% in the low group indicated that they would go to the same institution again (figure 4.4).

The third and final outcome we examined was those who answered no to the question “have you ever seriously considered withdrawing from your degree programme?”. As presented in figure 4.5, among those in the high quality of interaction group, 70.1% reported that they had not seriously considered withdrawing, while among the low quality of interaction group significantly less (42.0%) said they had not considered this.

While there are many factors behind why a student may consider withdrawing from their programme, who may wish to choose a different institution if they were to start over, or who’s overall evaluation of their institution is low, these results suggest that there is a significantly lower likelihood of reporting this among those students who rated the quality of interactions as high (compared to those who rate this as either low or medium), and furthermore the differences among the groups are large. The next section considers the role of other background factors in these findings.

The section above found that those students who rated the quality of interactions with staff and other students at their institution as high were significantly more likely to report that

1.  their overall evaluation of the institution was high,
2. that they would, if they started over, attend the same institution again, and
3. they had not considered withdrawal from their degree programme.

By modelling this relationship to include other factors, we consider whether these findings may be at least somewhat explained by differences in background factors among students, such as gender, age group, field of study, or the institution type that the student attended, or whether the findings are independent of these other factors.

Appendix 5 provides the fully adjusted logistic regression and coefficients for the three outcome measures modelled, reported in odds ratios[22]. For each of the three outcomes, the explanatory factor of quality of interactions was statistically significant after adjusting for these background factors.

[22] An odds ratio is a measure of association between an exposure and an outcome. The odds ratio represents the odds that an outcome will occur given a particular exposure, compared to the odds of the outcome occurring in the absence of that exposure (Szumilas, 2010). For example, in the present analysis the odds of the outcome of having a high overall  evaluation of the HE institution occurring among students reporting a high quality of interaction compared to it occurring among students reporting a low quality of interaction.

Figure 4.6 illustrates the results of the binary logistic regression model for the first outcome, positive overall evaluation of the institution, both before adjusting for background factors (in red) and after adjusting for these (in blue), and including confidence intervals at the 95% level. Where confidence intervals do not cross the threshold of an odds ratio of one, there is a statistically significant difference between the groups.

Before adjustment, those in the high quality of interaction group had a 17.2 times higher odds of having a positive overall evaluation than those in the low group, while those in the medium interaction group had a 3.3 times higher odds compared to the low group. After adjusting for gender, age group, field of study and institution type, the results were similar; the high quality of interaction group had a 17.5 times higher odds compared to the low group, while the medium group had 3.4 times higher odds of having a positive overall evaluation compared to the low group.

These results suggest the differences between the quality of interaction groups in the likelihood of having a positive overall evaluation of these institution was not explained by these background factors.

Figure 4.7 shows the results of the binary logistic regression model for the second outcome, affirming they would go to the same institution if they had their time over, both before and after adjustment for background factors, and including confidence intervals at the 95% level.

Before adjustment, the high quality of interaction group had 11 times higher odds of responding that they would choose the same institution again compared to those in the low group, while those in the medium group had a 3.5 times higher odds of doing so compared to the low group. Once again, the results are not substantially altered when the background factors are controlled for; the high group in this instance had an 11.3 times higher odds and the medium group a 3.6 times higher odds of responding that they would chose the same institution again, compared to the low quality of interaction group.

Figure 4.8 shows the modelled results for the third outcome, never seriously considered withdrawing from their programme. Before adjusting for gender, age group, field of study and institution type, the high quality of interaction group had a 3.2 times higher odds of not seriously considering withdrawal compared to the low group while the medium group had a 1.9 times higher odds than the low group. After taking account of the background factors, the model point estimates were unaltered.