Thursday, 28 February 2013

Triangulation in Social Sciences Research

By Mariya Razzaghian*

1. What is triangulation in research?

Triangulation is the combination of two or more theoretical perspective, methodological approaches, analytical methods, data sources, or investigators within the same study. Denzin (1970) has called mixing methods as a 'triangulation perspective'. 

2. Why we need triangulation?

Researchers use triangulation for two reasons mainly:
  1. The use of a single method may not allow the researcher to address all aspects of the research questions.
  2. Combining research methods increases the validity of the research findings because using a variety of methods means that one serves as a check on the other.
Since many researchers now recognize that there are differences in social science and natural science and that the interpretations that human beings have of their social behavior and the external world are important factors that need to be kept in mind, they now appreciate that both quantitative and qualitative methods have a role to play in social science. 

3. Types of Triangulation

There can be many types of triangulations. However, for the sake of parsimony  we shall focus on the four widely recognized types of triangulations here, i.e. 
  1. methodological triangulation 
  2. investigator triangulation, 
  3. theoretical triangulation, 
  4. data analysis triangulation.

3.1. Methodological Triangulation

This type of triangulation has also been called as mixed methods, multi-methods, and methods triangulation. By using multiple methods, the research strives to decrease the deficiencies and biases that stem from any single method creating the potential for counterbalancing the flaws or weaknesses of one method with the the strength of the other. Methodological triangulation can be further divided into two types, i.e. withing method and between methods triangulation.

3.1.1 With-in-method triangulation 

This is done at the data level and uses two different tools for collecting data. While, this type of triangulation has narrow scope, it provides alternative ways for measuring a variable.

3.1.2 Between methods triangulation 

we, may also call this across methods triangulation. It is done at the design level and uses both quantitative as well as qualitative methods.  The approaches are merged at the level of interpretation. 

3.2 Investigator Triangulation

It involves using more than one observer, interviewer, coder or data analyst in the study. Confirmation of data among investigators without prior decision or collaboration with one another lends greater credibility to the observation. The purpose of multiple investigator is to reduce the potential of bias in selecting, analyzing, coding, or gathering of data, as well as to increase the internal validity of the findings. Having more than one investigator also has the potential for keeping the team honest which increase the credibility of findings. Analysis of data (particularly quantitative data) by multiple analysts serves not only to amplify the results and increase validity but it also adds to the reliability.

Verification of the the interpretation of data (quantitative and qualitative  by more than one research can increase the value of the findings. Also, if the research is conducted by investigators who are equally skilled in their respective approaches can enhance the value of the study by contributing to the study from different view points.

Investigators equally apt in their respective research approaches can conduct the different phases of research allowing for both the investigators and methodological triangulation.

3.3 Theoretical Triangulation

This type involves using multiple theories and hypotheses to study a phenomenon. The use of multiple theories and hypotheses is done with the intention that the study can be conducted with multiple lenses and questions in mind in order to lend support to or refute the findings. 

The theoretical perspectives or hypotheses used in the study may be related to or they may have opposing views points, depending on what the researcher hopes to accomplish. The use of multiple theories provides the advantage that alternative explanation for a phenomenon can be decreased. Multiple theories also help to rule out competing hypotheses, prevent premature acceptance of plausible explanations, and increase confidence in constructing concepts in theory development.

3.4 Data Analysis Triangulation

Data analysis triangulation is the combination of two or more methods of analytical tools. These techniques can include different families of statistical tools such t-test, regression, anova, structural equation modelling in order to validate the data or determine similarities. 

*Mariya Razzaghian is a PhD student at the Institute of Management Sciences, Peshawar. Her area of specialization is HRM, with special focus on work place bullying and workaholism. 

Monday, 25 February 2013

Hypothesis Development: Lead-Lag in Stock Returns

The following hypotheses were developed during class discussion on the topic of whether some groups of stock lead others in their response to news shocks. The discussion took place in the Seminar in Finance Class, MS (Management) 2013. The paper for discussion was "Do Industries Lead Stock Market" [Download]

Hypotheses derived keeping in view slow diffusion of information and behavioral biases

H1: Firms with dispersed ownership will lead firms with concentrated ownership
Rationale: It is assumed that firms with concentrated ownership will have more information asymmetry as these firms will not share much information with the market [For literature review on this topic, you may read this paper].  In comparison, firms with dispersed ownership will have more information available. As news shocks hit the market, price discovery for firms with concentrated ownership will be difficult and hence they will react to the news with delay (similar rationale is available for non-synchronous trading hypothesis]

H2: Firms with less managerial ownership will lead firms with more managerial ownership
Rationale: Similar rationale can be developed for managerial ownership as for concentrated ownership. If managers have intention of expropriating wealth from minority shareholders, they will try to hide information from the market. With less information, share price discovery becomes difficult and hence such firms will react with delay to new information.

H3: Firms with higher trading volume will lead firms with lower trading volume.
Rationale: Firms with higher trading volume are more liquid and easy to price as they have more information available about them compared to firms with low trading volume. For more details, you can see paper Chordia and Swaminathan (2000) Download

H4: Firms with high beta stocks will lead firms with low beta stocks
Rationale: Since investors give more weight to loss than to a similar gain, stock with higher betas are expected to be affected by this behavioral bias. Hence in down markets investors are expected to sell these stocks quickly whereas lower beta stocks are expected to react with delay. For up market conditions, the theoretical predictions are not much clear about the two types of stocks.

H5: Small firms will lead large firms in down market trend.
Rationale: Though the above hypothesis seems to be in contradiction to the general finding of large stocks leading the small stocks, but based on the rational which we developed for the high beta stock it is plausible to expect that small firms will show quicker reaction to economic news in down markets (not because of information hypothesis but because of loss aversion hypothesis)

H6: Commodity market will lead stock market
Rationale: Since economic activity is primarily originated from commodity market, information from here will reach to the stock market with delay assuming limited cognitive ability (Kahneman, 1973; Nisbett and Ross, 1980) and limited participation Merton (1987) and Hong and Stein (1999)

H7: Firms in start of the supply chain will lead firms that are towards the end of the supply chain.
Rationale: Again the rationale is the slow diffusion of information. For details on this, you may read Menzly and Ozbas (2004)

H8: Mature firms will lead young firms.
Rationale: Again the rationale is the slow diffusion of information. Mature firms have more information about them compared to young firms.

H9: Stocks being followed by more analysts tend to lead stocks with no or less analysts.
Rationale: Institutional investors and analysts generate more analaysis and information about the firms in which they invest. Such firms are expected to have easier price discovery compared to other firms.

Chordia, Tarun, and Bhaskaran Swaminathan. "Trading volume and cross‐autocorrelations in stock returns." The Journal of Finance 55.2 (2002): 913-935.
Hong, H., Lim, T., Stein, J., 2000. Bad news travels slowly: Size, analyst coverage, and the profitability of momentum strategies. Journal of Finance 55, 265–295.
Hong, H., Stein, J., 1999. A unified theory of underreaction, momentum trading and overreaction in asset markets. Journal of Finance 54, 2143–2184.
Hong, H., Torous, W., Valkanov, R., 2002. Do industries lead the stock market? Gradual diffusion of information and cross-asset return predictability. FEN Working Paper /
Merton, R., 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance 42, 483–510.
Nisbett, R., Ross, L., 1980. Human Inference: Strategies and Shortcomings of Social Judgment. Prentice-Hall,New Jersey.
Kahneman, D., 1973. Attention and Effort. Prenctice-Hall, Englewood Cliffs, New Jersey.
Menzly, L., Ozbas, O., 2004. Cross-industry momentum. USC Working Paper.

Thursday, 21 February 2013

Welcome to OpenDoorsPK Blogging Area

I though several times to either follow or drop the idea of writing blogs. But ultimately I am here. The readers enthusiasm and response rate have always given me new spirit to delve into uncharted waters. In my busy schedules, I have to prioritize things. I would like to do the same for this blog as well. I would not waste my and your time in positing good-for-nothing material. Perhaps it would be more fruitful if you suggest topics of interest that can touch upon issues in my areas of expertise and at the same time are of interest to readers. Wish you happy reading.s

Attaullah Shah