WwA : Worldwide Attitudes
ISSN 1323-9589
Volume1995-04-03
© Copyright Clive Bean 1995. All rights reserved.
Public Attitudes to Party Leaders:
The Keating Factor
Clive Bean
Research School of Social Sciences
Australian National University
With John Howard having recently deposed Alexander Downer as leader of the Liberal
Party there has been renewed speculation about the extent to which Paul Keating's personal
style is an asset or a liability to the Labor government.
Arguments to the effect that Mr Keating's perceived arrogance and aggressiveness are
electorally harmful to his party have been fueled by the results of three recent
elections: (1) the Australian Capital Territory elections for the Legislative Assembly, in
which the Australian Labor Party lost office; (2) the New South Wales state elections,
which Labor won narrowly in circumstances under which some observers thought it should
have won handsomely; and, most notably, (3) the federal by-election for the seat of
Canberra, in which Labor suffered a massive swing against the party and lost to the
Liberals a seat generally regarded as safe for Labor in recent years.
Commentators in the media and even within his own party have suggested that these poor
showings are due, at least in part, to Mr Keating's unpopularity with the voting public.
The concern is that the prime minister's poor public image could cost the Labor Party
dearly at the next federal election. Mr Keating has stated that being popular with the
public is not the main aim of his political career. Unfortunately, however, past research
has shown that the popularity of the party leaders does matter in Australian elections and
that in a close election it can make a difference between winning and losing (Bean and
Kelley, 1988; Bean and Mughan, 1989).
In this sense, Mr Keating has a hard act to follow as his predecessor, Bob Hawke, was
in his prime one of the most popular political leaders in Australian history. Research has
shown that during the 1980s, Mr Hawke's popularity was a clear electoral asset to the
Labor Party. Some of this research has also shown that during that same period, when Mr
Keating was Treasurer, he in turn cost the Labor government some of the extra public
support that Mr Hawke won for it (Bean and Kelley, 1989). However, being Treasurer, which
entails having to announce and carry the responsibility for harsh economic decisions, is a
difficult role for any politician and it is likely that Mr Keating may be less unpopular
as prime minister.
Changing Attitudes to Keating and Howard
As a preliminary to examining Mr Keating's popularity at the time of the latest
National Social Science Survey (conducted in late 1994) and comparing it to that of Mr
Howard, we first look at how the popularity of the two has changed over the last eight or
so years (as shown by earlier NSSSs). The popularity of each leader is measured by a
device we call a "feeling thermometer", which asks respondents to rate each
leader on a scale ranging from 0 to 100, where 0 is a "very cold or
unfavourable" feeling and 100 is a "very warm or favourable" feeling, while
a score of 50 is neutral. We can then express the leader's overall popularity as the
average (or mean) score he receives on the scale from a representative sample of the
Australian public.
The accompanying chart shows that neither Mr Keating nor Mr Howard have ever been
terribly popular with the public, neither having achieved a rating as high as 50 (the
midpoint) in any of our surveys. (Mr Hawke, by contrast, scored over 60 in his heyday.) Mr
Keating's rating hovered around 40 during the late 1980s. After he took over as prime
minister it rose, being 45 around the time of the 1993 election, but by late 1994 it had
fallen back again, to 42.
When he was leader of the opposition, in the mid-1980s, Mr Howard also rated around 40
points on the feeling thermometer scale. Interestingly, Mr Howard became more popular
after he lost the Liberal leadership to Andrew Peacock, his rating rising to 46 points in
late 1989 and sitting at 48 points in 1990 and 1993. His popularity, too, had slid back
somewhat by late 1994, however. One might suggest that Mr Howard's periods of greatest
popularity have tended to coincide broadly with periods when he had a lower profile within
the ranks of the opposition.
Let us briefly consider more closely the latest data (from late 1994) on the ratings of
the two leaders (see Appendix A). As well as being slightly less popular overall than Mr
Howard, Mr Keating attracts a more polarised range of opinions. Thus, 22% rate him at the
lowest possible score of 0 on the thermometer scale, compared to only 12% for Mr Howard.
But at the same time, slightly more people actually give Mr Keating the highest possible
score of 100 than do the same for Mr Howard (4% versus 3%). And very few feel indifferent
about Mr Keating, rating him at the neutral point of 50 (only 8%, compared 14% for Mr
Howard).
Influences on Ratings of Keating and Howard
By applying multivariate analysis (multiple regression) to these ratings of the two
leaders we can see which of a range of factors influence the public's views of them (when
the effects of all the other variables in the equation are controlled for). The list of
possibilities includes a large number of socio-demographic factors plus salient political
and partisan attitudes (see Appendix B).
What we see is that by far the most prominent influences shaping ratings of the party
leaders are other political attitudes. So, for example, conservative party identifiers
(people who generally tend to support either the Liberal or National Party) are naturally
enough much less likely to rate Mr Keating highly than Labor Party identifiers. By
contrast, people who favour trade unions are much more inclined to like Mr Keating than
anti-unionists. Interestingly, however, those who favour big business in Australia also
tend to be pro-Keating to some extent, perhaps a reflection of the success of Mr Keating
and the Labor government in making policies that are attractive to the business sector.
A number of socio-demographic factors also have small, but statistically significant,
influences on attitudes to Mr Keating. For example, men rate the prime minister more
highly than women, high income earners find him more attractive than low income earners,
Victorians rate him higher than those in other parts of Australia (even though that is not
his home state) and British-born immigrants tend to be slightly anti Mr Keating (possibly
because of his well publicised views on the monarchy). Finally, it is intriguing to note
that, net of all the other influences in the equation, trade union members are likely to
rate Keating slightly lower than non-members, indicating perhaps some disquiet among this
traditional group of Labor supporters with either Mr Keating's style or his policies.
The public's views of Mr Howard are less well structured overall, except for the fact
that age has a very substantial impact. The opposition leader is much more popular with
older citizens than with the young. After that again it is partisanship (general support
for either major party grouping) and attitudes towards big business and trade unions that
have the largest roles in shaping views of Mr Howard. Understandably, partisanship and
attitudes towards unions work in the opposite direction for Mr Howard to the way they
effect views of Mr Keating (and the size of their effects are substantially lower), but
feelings about big business (traditionally a natural ally of the Liberal Party) have a
positive impact and a stronger influence than for the prime minister.
Among the socio-demographic influences, the overseas born are less likely than the
Australian born to view Mr Howard positively (presumably reflecting views on immigration
expressed by Mr Howard in the late 1980s that he has become well known for, views that he
has attempted to distance himself from in more recent times), while residents of New South
Wales (his home state) and the better educated find him more attractive than others.
The Electoral Impact of Leadership
Is Mr Keating still the electoral liability to his party that he appeared to be back
when he was Treasurer in the 1980s? The answer to this question has two components: his
level of popularity, relative to that of his party, and the level of his impact on
individual electoral choice. As we have already seen, Mr Keating's current level of
popularity (in late 1994) is 42.5 on the feeling thermometer scale. The equivalent rating
for the Labor Party is 46.8. Thus Mr Keating is some 4.3 points less popular than his
party and this means that any impact he has on the vote will work against the Labor Party.
We estimate Mr Keating's net impact on individual voting behaviour by predicting vote
choice with an equation that controls for all the socio-demographic and attitudinal
variables used earlier to predict attitudes to the leaders and also includes attitudes to
Mr Keating and Mr Downer (who, of course, was still Liberal leader at the time of the
survey). By far the major impact on vote choice is the generalised identification with one
or other of the major party groups, mentioned above. However, though much less important
than partisanship, attitudes to both major party leaders also have a sizable effect. (No
other variable in the equation has more than a very marginal impact.)
Voters' views on Mr Keating have a much larger influence on their voting behaviour than
do attitudes towards Mr Downer (and the same would almost certainly be true currently with
Mr Howard as Liberal leader). Voters who are most favourable to Mr Keating are 36% more
likely to vote for the Labor Party (rather than the Liberal or National parties) than
those who most dislike him. Voters who most favour the leader of the opposition are 23%
more likely to vote for the Liberal or National parties than those most against him.
When we combine the 36% impact of Mr Keating with the fact that he is rated 4.3 points
behind his party on the thermometer scale, a simple piece of arithmetic (36% time 4.3)
tells us that Mr Keating costs the Labor Party about 1.5% of votes. In other words, if Mr
Keating was as popular as his party, Labor would stand to gain another 1.5% of votes.
Although 1.5% perhaps does not sound very much, in a close election every factor that
sways the vote a little bit one way or the other matters. Mr Keating tends to sway the
vote a little bit away from Labor.
Since the time of this survey, while there is no reason to assume that the prime
minister's impact on voters would have changed, it is likely that if anything his level of
popularity may have fallen further. Thus, his liability to the Labor Party in this sense
may be greater now. Of course, it is possible that Mr Keating could turn that around
before the next election.
Finally, it is worth reflecting that while a party leader's level of public popularity
is an important electoral consideration, there are other ways in which politicians can
influence the electorate. These include setting and directing the agenda of public debate
so that the issues work in favour of the leader and his party. As evidenced by the
campaign Mr Keating waged in 1993 against the Fightback! program and the Goods and
Services Tax, no one in Australian politics is better equipped to do that than the prime
minister.
References
Bean, Clive and Jonathan Kelley. 1988. "Partisan Stability and Short-term Change
in the 1987 Federal Election: Evidence from the NSSS Panel Survey." Politics
23(2):80-94.
Bean, Clive and Jonathan Kelley. 1989. "Keating Would Cost ALP Votes." National
Social Science Survey Report 1(4):2-3.
Bean, Clive and Anthony Mughan. 1989. "Leadership Effects in Parliamentary
Elections in Australia and Britain." American Political Science Review
83(4):1165-79.
Appendix A: Frequencies
TKEATING p3q1a-Thermometer- Keating
Valid Cum
Value Label Value Frequency Percent Percent Percent
Very cold 0 282 22.1 22.3 22.3
1 1 .1 .1 22.4
2 3 .2 .2 22.7
3 1 .1 .1 22.7
5 2 .2 .2 22.9
6 1 .1 .1 23.0
8 2 .2 .2 23.1
9 1 .1 .1 23.2
10 17 1.3 1.3 24.6
15 117 9.2 9.3 33.8
20 10 .8 .8 34.6
25 5 .4 .4 35.0
30 106 8.3 8.4 43.4
35 2 .2 .2 43.6
38 1 .1 .1 43.7
40 79 6.2 6.3 49.9
45 1 .1 .1 50.0
50 105 8.2 8.3 58.3
52 1 .1 .1 58.4
55 4 .3 .3 58.7
60 142 11.1 11.3 70.0
65 5 .4 .4 70.4
70 136 10.7 10.8 81.1
75 11 .9 .9 82.0
80 20 1.6 1.6 83.6
85 139 10.9 11.0 94.6
89 1 .1 .1 94.7
90 10 .8 .8 95.5
95 3 .2 .2 95.7
Very warm 100 54 4.2 4.3 100.0
. 13 1.0 Missing
------- ------- -------
Total 1275 100.0 100.0
Mean 42.468 Std dev 32.272
Valid cases 1262 Missing cases 13
THOWARD p3q1f-Thermometer- Howard
Valid Cum
Value Label Value Frequency Percent Percent Percent
Very cold 0 143 11.2 11.6 11.6
1 5 .4 .4 12.0
2 3 .2 .2 12.2
3 2 .2 .2 12.4
5 6 .5 .5 12.8
10 25 2.0 2.0 14.9
15 127 10.0 10.3 25.1
20 14 1.1 1.1 26.3
22 1 .1 .1 26.3
25 9 .7 .7 27.1
30 110 8.6 8.9 35.9
35 4 .3 .3 36.3
36 1 .1 .1 36.3
40 151 11.8 12.2 48.5
45 7 .5 .6 49.1
50 170 13.3 13.7 62.8
51 1 .1 .1 62.9
55 8 .6 .6 63.6
60 177 13.9 14.3 77.9
65 4 .3 .3 78.2
70 126 9.9 10.2 88.4
75 7 .5 .6 88.9
80 19 1.5 1.5 90.5
85 73 5.7 5.9 96.4
90 8 .6 .6 97.0
Very warm 100 37 2.9 3.0 100.0
. 37 2.9 Missing
------- ------- -------
Total 1275 100.0 100.0
Mean 43.555 Std dev 27.002
Valid cases 1238 Missing cases 37
Appendix B: Regression Results
Variables in the analysis
Mean Std Devia Cases Label
MALE .540 .499 1255 Male=1
AGE 46.268 15.214 1248 Age in years
URBAN 73.629 30.393 1196 Urban/rural
EDUC 11.873 2.999 1275 Education (years)
GOVT .189 .392 1275 Works for government
BLUEC .301 .459 1022 Blue collar
FARM .033 .179 1022 Farm=1
FAMINC 42157.559 31770.481 1098 Family income (rough version)
TUMEMB .259 .438 1275 Trade union member
CATH .234 .423 1275 Catholic=1
NOREL .266 .442 1275 No religion=1
BORNPOM .093 .291 1275 Born in England
BORNMED .020 .139 1275 Born in Mediterranean nations
BORNOTH .124 .330 1275 Born in other places
NSW .287 .453 1275 NSW=1
VIC .264 .441 1275 Victoria=1
ACT .088 .283 1275 ACT=1
PNTPARTY .483 .395 1275 Parents political party (Conservative=1)
CONSPTY .475 .458 1275 Party ID (Conservative=1)
PROUNION 39.510 19.586 1274 Pro-trade union (0 to 100)
TBIGBUS 51.382 21.678 1220 p4q3d-Thermometer- Australian Big Busine
TKEATING 42.468 32.272 1262 p3q1a-Thermometer- Keating
THOWARD 43.555 27.002 1238 p3q1f-Thermometer- Howard
TDOWNER 39.340 25.473 1247 p3q1b-Thermometer- Downer
TLIBS 43.841 27.392 1254 p3q2a-Thermometer- The Liberal Party
TLABOR 46.833 28.458 1256 p3q2b-Thermometer- The Labor Party
VOTECONS .494 .466 1235 p6q3-If House of Reps election tomorrow
Regression equations
Equation Number 1 Dependent Variable.. TKEATING p3q1a-Thermometer- Keatin
R Square .48036
Variable B SE B Beta T Sig T
TBIGBUS .168491 .036481 .113182 4.619 .0000
FAMINC 6.39673E-05 2.6236E-05 .062974 2.438 .0150
FARM -4.699520 4.754979 -.026128 -.988 .3233
VIC 4.577517 2.008940 .062573 2.279 .0229
BORNMED 10.181784 5.666337 .043761 1.797 .0727
BORNOTH 4.558331 2.462644 .046559 1.851 .0645
MALE 4.418965 1.605952 .068270 2.752 .0060
AGE -.017863 .056666 -.008422 -.315 .7527
PROUNION .471964 .048512 .286440 9.729 .0000
NOREL 2.961441 1.919086 .040558 1.543 .1231
BORNPOM -5.787085 2.737290 -.052186 -2.114 .0348
ACT -2.624615 2.948175 -.023031 -.890 .3736
GOVT 1.871976 2.203106 .022720 .850 .3957
PNTPARTY -2.092368 2.191702 -.025629 -.955 .3400
BLUEC -1.851103 1.918416 -.026333 -.965 .3348
CATH 2.819751 2.002810 .036992 1.408 .1595
URBAN .032395 .028116 .030509 1.152 .2495
TUMEMB -4.405460 2.002662 -.059814 -2.200 .0281
NSW 1.485084 1.953543 .020826 .760 .4473
EDUC .056665 .320550 .005266 .177 .8597
CONSPTY -32.757717 2.191268 -.464863 -14.949 .0000
(Constant) 22.839582 6.584674 3.469 .0005
Equation Number 2 Dependent Variable.. THOWARD p3q1f-Thermometer- Howard
R Square .24719
Variable B SE B Beta T Sig T
TBIGBUS .232089 .036740 .186327 6.317 .0000
FAMINC -5.10691E-06 2.6422E-05 -.006009 -.193 .8468
FARM 1.314147 4.788676 .008732 .274 .7838
VIC 1.687666 2.023177 .027572 .834 .4044
BORNMED -7.953469 5.706493 -.040855 -1.394 .1637
BORNOTH -8.290447 2.480096 -.101203 -3.343 .0009
MALE -.652273 1.617333 -.012044 -.403 .6868
AGE .452817 .057067 .255141 7.935 .0000
PROUNION -.195247 .048855 -.141622 -3.996 .0001
NOREL -2.126529 1.932686 -.034807 -1.100 .2715
BORNPOM -4.120311 2.756689 -.044406 -1.495 .1354
ACT -.307163 2.969068 -.003221 -.103 .9176
GOVT .299316 2.218718 .004342 .135 .8927
PNTPARTY 3.990847 2.207234 .058423 1.808 .0709
BLUEC -2.010538 1.932011 -.034182 -1.041 .2983
CATH -2.226754 2.017004 -.034913 -1.104 .2699
URBAN .013553 .028315 .015254 .479 .6323
TUMEMB .417355 2.016854 .006772 .207 .8361
NSW 4.481401 1.967387 .075110 2.278 .0230
EDUC .759800 .322822 .084388 2.354 .0188
CONSPTY 11.295302 2.206797 .191571 5.118 .0000
(Constant) 2.995942 6.631338 .452 .6515
Equation Number 3 Dependent Variable.. VOTECONS p6q3-If House of Reps ele
R Square .77088
Variable B SE B Beta T Sig T
TDOWNER .002336 3.4096E-04 .127727 6.853 .0000
FAMINC 1.12090E-07 2.5266E-07 .007642 .444 .6574
VIC -.007667 .019339 -.007259 -.396 .6918
BORNPOM .003842 .026339 .002399 .146 .8841
BORNMED .056245 .054485 .016742 1.032 .3022
FARM -.071008 .045668 -.027341 -1.555 .1203
NOREL -.007661 .018468 -.007266 -.415 .6784
MALE -.007606 .015539 -.008138 -.489 .6246
BORNOTH .006299 .023708 .004456 .266 .7905
AGE 2.18016E-04 5.4496E-04 .007118 .400 .6892
ACT .020562 .028322 .012496 .726 .4680
GOVT .004663 .021173 .003920 .220 .8257
TBIGBUS 5.68270E-04 3.6295E-04 .026437 1.566 .1178
PNTPARTY -.015286 .021054 -.012967 -.726 .4680
BLUEC -.029743 .018447 -.029303 -1.612 .1072
PROUNION -6.46156E-04 4.8962E-04 -.027159 -1.320 .1873
CATH .019385 .019271 .017612 1.006 .3147
URBAN -4.83152E-04 2.7010E-04 -.031513 -1.789 .0740
TUMEMB .039589 .019276 .037226 2.054 .0403
NSW -.003080 .018763 -.002992 -.164 .8696
EDUC .001969 .003077 .012674 .640 .5224
TKEATING -.003638 3.2055E-04 -.251962 -11.350 .0000
CONSPTY .630941 .024164 .620086 26.111 .0000
(Constant) .258244 .064180 4.024 .0001
Visitors since 14 September 1998: 