Establishing Causality is Difficult Whether Conclusions have been Derived Deductively or Inductively (Term Paper Sample)

Business Research Methods
Name:
Industrial Affiliation:
Business Research Methods
1. Establishing causality is difficult whether conclusions have been derived deductively or inductively.
I) Explain and elaborate on the implications of this statement.
To meet the standards of causation, it is prescribed that a certain cause must only result to a specific result. For instance, if A is equal to B, then we expect to find only B in instances where we find A. In other words, if we can find B and only B in all instances where we find A, then we conclude that A and B are causally related. Thus, it is theoretically expected that for a causal relationship, a specific cause must be the only factor that leads to the implied result. However, in a practical aspect, it often seen that various causes and effect intertwine and are interdependent, thus it is difficult to implicitly conclude that the specified result is a cause of a specific cause whether you carry out your research deductively or inductively.
ii) Why is ascribing causality more difficult when conclusions have been reached through induction?
Attributing cause and effect is more difficult using inductive methods. First, inductive reasoning relies on the process of establishment of patterns or relations from the collected data. However, it is obvious that there are several causes to a specific result. Thus, in ascribing causality using inductive reasoning, it is more probable to come across several probable causes to a specific result or several probable results from a specific cause. Thus, the researcher is mainly boxed into determining casualty given ceteris peribus. Secondly, one is never hundred per cent when partaking in inductive reasoning. This is because inductive reasoning is limited to the data available at hand. Therefore, it is likely that a relation might exist that is not included in the data. For instance, most stock market crashes are due to flawed inductive reasoning. Finally, ascribing causality is more difficult in inductive reasoning due to bias. Every human is predisposed to suffer from confirmation bias. This is because specific trends exist which are expected. This might cause some difficulty in the determination of the relationship between cause and effect.
iii) Correlation does not imply causation. Illustrate this point with examples from business
The assumption that correlation implies causation is a fallacy better explained by its Latin terms ‘cum hoc ergo propter hoc’ which means that if a result occurs with this consequence, then it is assumed to be the cause or ‘pro hoc ergo propter hoc’ which assumes an effect if the result usually occurs after the specified action. However, this is no true. For example, it is observed that when a country’s debt rises above 90 per cent, GDP slows. Thus it has been falsely assumed that high debt causes a country’s growth rate to slow. This is a factor that has been proven to be false since increase in debt actually increase the money in circulation above the capacity of a country, hence paving way for more development. It is also observed that those who source capital for their own businesses are more likely to succeed than those who source capital from external sources. Thus, it is assumed that sourcing money from your income increases the chances of success of the business. However, it is the fact that those who source their money tend to be more dedicated to the business and tend to put more effort in management that leads to success of the business.
2.
a) Identify and give a detailed discussion the five issues (questions) that a researcher needs to answer when designing a sample for a study.
How will the sampling error and non-sampling error be balanced? In retrospect, the researcher must determine how to balance the errors due to sampling. The researcher must first evaluate what errors can occur if he uses a sample of the population. This should be considered in design where the study should be designed such that these errors are minimized. For instance, using cluster sampling in the case that the population characteristics varies greatly. In non-sampling error, the designer should ensure that methods used cause as little error as possible.
What is the relevant population? In this case, the population being taken into account must be considered. This will inform the designer on the methods of research to be chosen in order to access as much data as possible.
What is the suitable margin of error to be achieved? In the case, the researcher should consider the errors that the research is allowing to include during the research. This ensures that the methods of data collection and analysis are chosen such that they give allowance to errors likely to be present.
What is the relevant sampling unit? In this case, one should factor in the population and consider the most representative single unit. For instance, in a farming community in a rural area, the most appropriate unit would be in the urban center with the major market since all of the population is represented.
What is the appropriate sampling frame? Here, it is important to choose the frame that is most representative of the unit being researched. For instance, a directory would not be an appropriate sampling frame for a research entailing single unit shops.
b) Give a detailed description and comparison of the five probability sampling designs (simple random sampling, systematic, stratified, and cluster) and clearly discussing their advantages and disadvantages
In simple random sampling, it is assumed that each of the population has an equal probability of being selected. Sampling is done with or without replacement where units are selected at random and each unit given a number. The population characteristics are assumed to be uniform. This sampling is especially advantageous in that it is very economical and time saving. It is also simple to carry out thus requires little skill in execution. However, the method is disadvantageous in that it assumes that the population is uniform and if that was not to be the case, the errors accrued would be tremendous.
Systematic sampling (interval random sampling) involve a strategic design to ensure equality in the sampling process. In this case, an interval is presented after which a unit is picked. This ensures that there is equal representation of the whole population. This sampling is especially useful in that it ensures equal representation of units in sampling. However, it can be disadvantageous in that it might facilitate insufficient data collection.
Stratified sampling is where data is first grouped according to various characteristics and then samples are chosen from the group. It can be divided into proportionate stratified (the strata sample size are made proportional to the strata population size) and disproportionate stratified (varying sampling is used). This method is especially advantageous in that it ensures that the quality is high and also ensures that most of the population is represented. However, it has a downfall in that it makes the researcher susceptible to pro bias incase inductive logic is used.
Cluster (Area) sampling is where the sample is first grouped according to the most basic characteristic to be considered. Sampling is then made according to all the stages of the cluster. The advantage of this is that it enables comparative analysis can be done after the research to determine extra...