WEBVTT

00:00:01.920 --> 00:00:04.900
Eisenhardt's approach
to multiple case studies

00:00:04.900 --> 00:00:07.930
has a very structured way
of presenting evidence.

00:00:07.930 --> 00:00:08.850
Let's take a look at

00:00:08.850 --> 00:00:11.423
how Eisenhardt presents
evidence in her research.

00:00:12.660 --> 00:00:14.640
We'll take a look at the other purposes

00:00:14.640 --> 00:00:15.790
in one of these paper,

00:00:15.790 --> 00:00:17.850
Eisenhardt and Burgeois, 1988.

00:00:17.850 --> 00:00:20.600
The proposition is that the
greater decentralization

00:00:20.600 --> 00:00:23.640
of power to a chief executive,

00:00:23.640 --> 00:00:26.530
the greater use of politics
within top management team.

00:00:26.530 --> 00:00:27.907
So this is just one proposition

00:00:27.907 --> 00:00:30.850
but is the most important
proposition in the paper

00:00:30.850 --> 00:00:32.750
and the paper also contained

00:00:32.750 --> 00:00:36.500
many other propositions about how power

00:00:36.500 --> 00:00:39.690
and centralization of politics
influence performance.

00:00:39.690 --> 00:00:41.160
But let's take a look at simply

00:00:41.160 --> 00:00:43.500
this power central assistant and politics.

00:00:43.500 --> 00:00:47.560
So our first, how do you
present qualitative evidence?

00:00:47.560 --> 00:00:51.030
There is, there are basically
are three different levels

00:00:51.030 --> 00:00:53.350
on which you can present evidence.

00:00:53.350 --> 00:00:57.450
When a beginner or qualitative
researcher enters the field,

00:00:57.450 --> 00:01:00.350
we quite often see this kind of reporting.

00:01:00.350 --> 00:01:02.390
So interviews were done.

00:01:02.390 --> 00:01:04.790
I conclude that something happens

00:01:04.790 --> 00:01:06.730
and this kind of reporting

00:01:06.730 --> 00:01:11.730
basically relies on the
competence of the researcher.

00:01:12.220 --> 00:01:15.800
So there is the claim is valid

00:01:15.800 --> 00:01:18.720
because you claim that
you are a good research

00:01:18.720 --> 00:01:21.270
and that's not a very good way

00:01:21.270 --> 00:01:23.640
of presenting evidence or making claims.

00:01:23.640 --> 00:01:25.500
So just saying that
you went to the company

00:01:25.500 --> 00:01:26.760
and you interviewed people,

00:01:26.760 --> 00:01:29.670
and then saying that
based on those interviews,

00:01:29.670 --> 00:01:33.120
you came to the conclusion
that power was centralized.

00:01:33.120 --> 00:01:36.110
How would we know that you
made the right conclusion?

00:01:36.110 --> 00:01:38.440
So researchers opinion,

00:01:38.440 --> 00:01:40.830
even if it's based on analysis

00:01:40.830 --> 00:01:44.493
is not the ideal way
of presenting evidence.

00:01:45.960 --> 00:01:48.920
So that's not really a good way.

00:01:48.920 --> 00:01:53.040
The second way is presenting
the informants opinions

00:01:53.040 --> 00:01:56.960
or information months
are reporting as results.

00:01:56.960 --> 00:01:59.980
So you might go to a
company, interview the CEO

00:01:59.980 --> 00:02:01.930
and asked the CEO if power is centralized,

00:02:01.930 --> 00:02:03.060
the CEO says, yes.

00:02:03.060 --> 00:02:07.000
And then you report that there
are the power is centralized

00:02:07.000 --> 00:02:08.640
because the CEO said so.

00:02:08.640 --> 00:02:12.370
And this is simply the
same as the one before

00:02:12.370 --> 00:02:14.420
but instead of reporting your own opinion,

00:02:14.420 --> 00:02:16.490
you report somebody else's opinion.

00:02:16.490 --> 00:02:19.320
How do we know that the
CEO is actually correct?

00:02:19.320 --> 00:02:21.240
We really don't.

00:02:21.240 --> 00:02:23.590
So that's not an ideal approach either.

00:02:23.590 --> 00:02:26.600
The better approach is to present evidence

00:02:26.600 --> 00:02:30.507
so what exactly did the interviewee say?

00:02:31.937 --> 00:02:35.690
And when were multiple
interviews that describe things

00:02:35.690 --> 00:02:39.820
that are associated with
power centralization

00:02:39.820 --> 00:02:40.950
then we can show that,

00:02:40.950 --> 00:02:44.660
okay, this is what the informants told us

00:02:44.660 --> 00:02:48.510
and that allowed us to
come up with the conclusion

00:02:48.510 --> 00:02:50.920
that the power must be centralized.

00:02:50.920 --> 00:02:53.820
And this is the ideal approach
for presenting evidence,

00:02:53.820 --> 00:02:55.620
because it allows your reader

00:02:55.620 --> 00:02:59.020
to also do the analysis themselves.

00:02:59.020 --> 00:03:03.160
So would the reader based
on hearing what you heard

00:03:03.160 --> 00:03:04.580
make the same conclusion,

00:03:04.580 --> 00:03:07.150
and that gives the reader the chance

00:03:07.150 --> 00:03:08.390
to form their own opinion

00:03:08.390 --> 00:03:10.610
of whether you analyze the data correctly.

00:03:10.610 --> 00:03:13.270
This is called transparency of reporting,

00:03:13.270 --> 00:03:16.350
and it is very important
in qualitative research

00:03:16.350 --> 00:03:18.550
because qualitative
research is more subjective

00:03:18.550 --> 00:03:19.900
than quantitative research.

00:03:20.870 --> 00:03:21.930
Let's take a look at

00:03:21.930 --> 00:03:25.093
another important
principle, triangulation.

00:03:25.950 --> 00:03:28.640
So quite often you'll
see something like that.

00:03:28.640 --> 00:03:30.690
So one executive told us

00:03:30.690 --> 00:03:32.460
that Geoff runs the show.

00:03:32.460 --> 00:03:34.860
Geoff is like God and so on.

00:03:34.860 --> 00:03:36.770
And this is not a democracy.

00:03:36.770 --> 00:03:39.230
So this is what the informants told us,

00:03:39.230 --> 00:03:42.360
but simply what the informants told us

00:03:42.360 --> 00:03:43.890
is often not sufficient.

00:03:43.890 --> 00:03:46.170
So we want to all come up with

00:03:46.170 --> 00:03:48.240
other sources of evidence as well.

00:03:48.240 --> 00:03:51.780
So we might add some quantitative data.

00:03:51.780 --> 00:03:54.890
So for example, we have a survey,

00:03:54.890 --> 00:03:57.500
a form that the informants complete,

00:03:57.500 --> 00:04:00.080
that measures power
centralization of some,

00:04:00.080 --> 00:04:01.610
with some kind of scale.

00:04:01.610 --> 00:04:05.090
And we have high scores
for this organization.

00:04:05.090 --> 00:04:09.330
We might also have an objective fact,

00:04:09.330 --> 00:04:11.360
for example, meeting minutes

00:04:11.360 --> 00:04:13.770
that the CEO was the decision maker

00:04:13.770 --> 00:04:16.770
in multiple different functional areas.

00:04:16.770 --> 00:04:19.970
So based on these different
sources of information,

00:04:19.970 --> 00:04:23.330
quantitative data,
information from the minutes

00:04:23.330 --> 00:04:24.700
of who makes the decisions,

00:04:24.700 --> 00:04:26.440
and then how if there are people

00:04:26.440 --> 00:04:28.270
in the organization perceive things,

00:04:28.270 --> 00:04:31.260
we can conclude that power is centralized

00:04:31.260 --> 00:04:33.360
to the CEO, Geoff here.

00:04:33.360 --> 00:04:36.720
So multiple sources
using different measures.

00:04:36.720 --> 00:04:39.420
And this is actually
what the paper reports.

00:04:39.420 --> 00:04:41.780
So they report that anecdotes

00:04:41.780 --> 00:04:44.820
from interviews corroborated
the quantitative data.

00:04:44.820 --> 00:04:48.270
Then, what you need to do
is demonstrate association.

00:04:48.270 --> 00:04:51.590
Eisenhardt approach uses
tables for this purpose.

00:04:51.590 --> 00:04:52.660
So what she does

00:04:52.660 --> 00:04:55.100
is that she tabulates different evidence

00:04:55.100 --> 00:04:56.690
of power centralization.

00:04:56.690 --> 00:04:58.945
So we have different aspects that,

00:04:58.945 --> 00:05:00.250
different things that we could observe

00:05:00.250 --> 00:05:01.720
if power is centralized,

00:05:01.720 --> 00:05:04.230
we have CEO description, CEO power score,

00:05:04.230 --> 00:05:07.100
CEO power distance,
CEO dominated function,

00:05:07.100 --> 00:05:10.350
decision styles and then examples

00:05:10.350 --> 00:05:15.350
of how people explain the
CEO's power or use of power.

00:05:15.600 --> 00:05:17.100
Then we infer that power

00:05:17.100 --> 00:05:20.070
is centralized in the first four cases

00:05:20.070 --> 00:05:22.100
and power is not centralized

00:05:22.100 --> 00:05:24.250
in the last four cases.

00:05:24.250 --> 00:05:28.300
So we quantify the qualitative evidence

00:05:28.300 --> 00:05:31.930
and we also combine it with
some quantitative data.

00:05:31.930 --> 00:05:34.210
So these, forming these tables

00:05:34.210 --> 00:05:36.620
is one of the first steps

00:05:36.620 --> 00:05:39.290
when you start writing
the results section,

00:05:39.290 --> 00:05:41.000
when you apply Eisenhardt approach

00:05:41.000 --> 00:05:42.353
to multiple case studies.

00:05:43.270 --> 00:05:45.580
Then you need to code another construct.

00:05:45.580 --> 00:05:48.380
So this is the use of politics.

00:05:48.380 --> 00:05:50.560
And then this article Eisenhardt coded

00:05:51.590 --> 00:05:53.410
these different tactics

00:05:53.410 --> 00:05:55.510
of different politic tactics

00:05:55.510 --> 00:05:57.230
and whether they found evidence

00:05:57.230 --> 00:05:59.670
of these tactics within the cases

00:05:59.670 --> 00:06:02.690
and if there were more
of these tactics applied,

00:06:02.690 --> 00:06:06.830
then are the use of power
of politics was more common

00:06:06.830 --> 00:06:09.990
than if there was just a few
of these tactics applied.

00:06:09.990 --> 00:06:12.700
And we can see that politics is practiced

00:06:12.700 --> 00:06:15.570
in these organizations, these three, four.

00:06:15.570 --> 00:06:16.990
Then three first,

00:06:16.990 --> 00:06:19.370
the fourth is kind of like maybe,

00:06:19.370 --> 00:06:21.750
the fifth is kind of like maybe as well.

00:06:21.750 --> 00:06:25.330
And then these final three are no.

00:06:25.330 --> 00:06:28.030
We can see here that
there's a clear association

00:06:28.030 --> 00:06:30.460
that where politics is practiced,

00:06:30.460 --> 00:06:32.140
power is also centralized.

00:06:32.140 --> 00:06:34.900
So there's clear evidence of association,

00:06:34.900 --> 00:06:36.810
but how do we then claim

00:06:36.810 --> 00:06:40.610
that this association is
actually causal relationship?

00:06:40.610 --> 00:06:44.680
We do so by explaining the causal process,

00:06:44.680 --> 00:06:47.620
and they explained the
course of process here.

00:06:47.620 --> 00:06:48.990
So they show first,

00:06:48.990 --> 00:06:52.660
say that there's evidence
of politics in table three.

00:06:52.660 --> 00:06:56.900
And then they start to analyze
why was politics applied?

00:06:56.900 --> 00:07:01.880
They found a couple of
facts from the data.

00:07:01.880 --> 00:07:05.350
So executive view politics as beneficial

00:07:05.350 --> 00:07:07.340
to have power over CEO.

00:07:07.340 --> 00:07:11.710
So executives view the
politics were useful.

00:07:11.710 --> 00:07:14.680
Executives are also
frustrated with the CEO

00:07:14.680 --> 00:07:17.390
because the CEO had
power centralized to him

00:07:17.390 --> 00:07:19.250
and the executives felt that

00:07:19.250 --> 00:07:22.370
they did not have enough power themselves.

00:07:22.370 --> 00:07:25.890
And finally politics
were seen as necessary

00:07:25.890 --> 00:07:27.700
to getting things done

00:07:27.700 --> 00:07:31.970
and getting things decided

00:07:31.970 --> 00:07:34.653
even if the CEO wanted
to decide otherwise?

00:07:35.930 --> 00:07:37.720
So, we have these facts;

00:07:37.720 --> 00:07:40.500
executive view politics as beneficial,

00:07:40.500 --> 00:07:42.780
executives are frustrated with the CEO

00:07:42.780 --> 00:07:47.110
and politics were seen as
necessary to get things done.

00:07:47.110 --> 00:07:50.070
Based on these facts and the association,

00:07:50.070 --> 00:07:51.760
we then write a theory,

00:07:51.760 --> 00:07:52.740
which could say that

00:07:52.740 --> 00:07:55.730
the centralization of power
to the CEO causes frustration

00:07:55.730 --> 00:07:59.190
in other executives because
the executives feels

00:07:59.190 --> 00:08:00.830
that they cannot influence the systems

00:08:00.830 --> 00:08:01.960
in the normal way.

00:08:01.960 --> 00:08:04.340
They feel that the politics
are the only alternative.

00:08:04.340 --> 00:08:06.210
By this way, centralization of power

00:08:06.210 --> 00:08:09.940
to the CEO causes other
executives to use politics.

00:08:09.940 --> 00:08:11.680
So this is an explanation

00:08:11.680 --> 00:08:15.240
of why the (indistinct) CEOs use politics

00:08:15.240 --> 00:08:18.810
and why the CEO power was the trigger

00:08:18.810 --> 00:08:20.560
for the use of politics.

00:08:20.560 --> 00:08:21.910
And this is basically

00:08:21.910 --> 00:08:26.203
how the proposition one
is justified in the paper.