Poaching the personal vote: How shadowing behaviour shapes constituent impressions

ABSTRACT

There are well-documented tensions between regional and constituency members of the Scottish Parliament (MSPs). Much of this friction stems from regional MSPs competing over constituency casework through shadowing. Although these so-called dual-mandate MSPs clearly see value in pursuing a personal vote, little is known about whether these efforts affect the perceptions of constituents. This article presents a survey of Scottish respondents to show that dual-mandate MSPs are better known and enjoy higher levels of approval than regional members who did not contest a constituency seat in the last election. We find dual-mandate MSPs enjoy similar levels of regard as constituency MSPs. Moreover, these advantages accrue even among constituents who do not share their MSP's partisan affiliation. We conclude by considering the normative impact of shadowing, which may generate unanticipated representational consequences for citizens difficult to remedy in a unicameral legislature.

KEYWORDS:

• Scottish Parliament
• representation
• personal vote
• mixed-member proportional
• electoral connection

Acknowledgements

The authors acknowledge the helpful advice and comments of Chris Carmen, Thomas Lundberg, and Mark Shephard. An earlier version of this project was presented at the Political Science Association's annual meeting held in Cardiff in March 2018.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 But see Stoffel (2014) who finds that the distribution of highway bypass projects by members of the Bundestag depends mostly on electoral vulnerability regardless of whether the MP is list, constituency, or dual-candidate.

2 Parties can also require MSPs to run on regional lists and in constituencies. According to a private conversation with one of the authors with an MSP on 30 May 2017, all Conservative candidates for the Scottish Parliament had to run for a constituency and regional seat in 2016. The Scottish Labour party had long banned dual candidacy, but eliminated the ban for the 2016 election. Tentative evidence suggests that positive contamination effects exist that make shadowing beneficial for parties. Of the 45 constituencies where a dual-mandate MSP ran in the 2016 Scottish Parliamentary elections, the dual-mandate candidate's party received a higher percentage of the list vote cast in 32 of those constituencies than the party achieved in the region.

3 Tensions between regional and constituency members seem largely absent in Germany. ‘Constituency representation in Germany is not an exclusive right … but rather a responsibility shared among members of different parties’ writes Lundberg (2007, 167).

4 Battle (2019) and Poole (2019) examine how dual-mandate members face different incentives when choosing committee assignments and the amount of attention paid to committee work.

5 The average respondent age was 50, the median income bracket was £35,000–£40,000, and 35.1% had some form of university education. Women comprised 51.9% of the sample. Roughly 51% of respondents reported being affiliated with the Scottish National Party, 25% with the Conservatives, 18% with Labour, 3% with the Greens, and 2% with the Liberal Democrats.

6 See online appendix A for the survey instrument.

7 Average time for survey completion was 19 minutes.

8 Online appendix C details how we ensure unbiased estimates of constituent perceptions.

9 We deleted observations for respondents living in Strathkelvin and Bearsden as we mistakenly listed the retired constituency MSP on the survey.

10 Respondents answering ‘Don't know’ were excluded.

11 Although the percentage of respondents who had met their representative or attended a surgery was relatively low, it always exceeded a total of 200 observations, eliminating the need for a rare events analysis (King and Zeng 2001).

12 With all other variables set to their means.

13 Evaluations are on 5-point scale, where one is the most negative response option, three represents a neutral midpoint, and five is the most positive response option. Predicted probabilities reflect the probability of selecting either the fourth or fifth response option. They were derived by collapsing the scale to 3-points and re-running the logistic regressions, with substantially similar results (available from authors).

14 See online Appendix G for a discussion on how a dual-mandate MSP's party and number of shadowing attempts affect constituency perceptions. We find that neither seem to matter.