Scientific Writing

Chapter 4: Hypotheses, Questions,

and Evidence

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

1

Prepared by:

• Sidra Hashimi
• Aulya Hamidi

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

2

Hypothesis

• Research is guided by specific question that needs answer
• Questions are based on some informal model are how things work or how they interact or how they behave
• Hypothesis is a statement of belief about how the object behaves
• Example: we examine the limits of speech recognition, ask whether web search can be used effectively by children, or predict how well a service will respond to increasing load
• How to proceed?
• Speculate the impact of our algorithm or approach, i.e., reduces computational cost
• Perform preliminary investigation to obtain your hypothesis
• Conduct appropriator experimentation to observe that the hypothesis is correct and provide evidence

​

​

​

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

3

��Hypothesis and Research Question��

• Hypothesis and research question should be specific, precise and unambiguous.
• Indicate the limits of the conclusion as well.
• Example: Suppose p-list are a well- known date structure used for a range of application, in particular as an in –memory search structure that is fast and compact. A scientist has developed a new data structure called the Q-list . Formal analysis has shown the two structures have the same asymptotic complexity in both space and time , but the scientist intuitively believes the Q-list to be superior in practice and has decided to demonstrate this by experiment.
• Ideas con not be know to be correct when they are first conceived ,it is intuition or plausibility that suggests them as worthy of consideration.

​

​

​

​

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

4

Hypothesis

• Not OK: Q-lists are superior to p-lists
• Applied to all applications, in all conditions, for all time?
• OK: As an in-memory search structure for large data sets, Q-lists are faster and more compact than p-lists

Writing for Computer Science

Introduction

Computer Science

Information Systems

5

�Hypothesis Testability�

• Hypothesis can be falsified. Vague statement cannot be tested or falsified.
• Not OK: list performance is comparable to P-list performance
• Not OK: Our proposed query language is relatively easy to learn
• As an in-memory search structure for large data sets, Q-lists are faster and more compact than p-lists.
• Claiming about a black box algorithm is invalid. Black box algorithm means that the algorithm is poorly understood.
• Experimentation is not something where results are excellent on a specific data.
• ­Result should be general that they could be applied to future data and also on theories or models
• Some researches fancy terminologies
• Using fancy terms wont give any value to your research. Ex., “network cache” is called as “local storage agent”

​

​

​

​

​

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

6

�Hypotheses�

• Hypotheses should not following a specific data
• “the algorithm worked on our data”
• “the algorithm was predicted to work on any data of this class, and this prediction has been confirmed on our data”
• Provide hypotheses first and then test
• Don’t fine–tune the experiment and/or the hypotheses based on the data you have
• Choose simple hypotheses is if two hypotheses are equally fit for the observation

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

7

Defending Hypothesis

• Defending the hypothesis is vital for any paper
• There should be argument that relates your hypothesis with the evidence
• The hypothesis “the new range searching method is faster than previous methods” might be supported by evidence “range search amongst n elements requires 2 log2 n + c comparisons”

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

8

�Defending Hypothesis (continued)�

• Always provide reasoning for objections that could be raised or those you did raise while forming the hypothesis
• Any objection that could not be refuted should also be there in the paper. but this could open door for further refinement or interpretation of results
• What if the hypothesis contradicts common belief?
• Verify whether all observations or scenarios in the common belief are considered or not
• Initially, be sceptical about your hypothesis; but, prove it using tests
• Taking too long to prove any hypothesis is just a waste of time
• If you have stronger intuition toward the hypothesis, then test is rigorously as well

Writing for Computer Science

Introduction

Computer Science

Information Systems

9

Convince The Reader

• Let us consider that you have proposed an algorithm to solve a problem
• Will the reader believe that algorithm is new?
• Careful and critical literature review is must; Give sufficient credit to any advancement in the literature
• Will the reader believe that the algorithm is sensible?
• Provide potential problem with appropriate
• Are the experiments convincing?
• Can the code and data be made public? is the data standard?

Writing for Computer Science

Introduction

Computer Science

Information Systems

10

Forms of Evidence

• Paper is about evidence and supporting explanations
• Good science expects objective evidence for hypothesis or question
• Evidence should convince the critical and skeptical reader
• Evidences to support hypothesis: proof, modelling, simulation, Experiment
• Proof:
• Formal argument that the hypothesis is true or false
• Proof can also have error

Writing for Computer Science

Introduction

Computer Science

Information Systems

11

​

Forms of Evidence (continued)�

• Modelling
• Mathematical description that corresponds to the hypothesis
• Simulation
• Implementation (in full or part) of the hypothesis
• Some approximations and /or omissions can be made while making the implementation
• Experiment
• Full test of hypothesis on real (at least highly realistic) data
• Experiment is doing it; Simulations is pretending to do it
• Limitation: Experiment confirms that hypothesis holds for a specific data set; Modeling and simulation generalizes the conclusion

Writing for Computer Science

Introduction

Computer Science

Information Systems

12

Use of Evidence

​

• Validation can be performed using different evidences
• For some proposed algorithm, a mathematical mode is designed. If this model is implemented and performance is predicted, is this an experiment?
• NO. this only confirms the model
• Which evidence to choose?
• Choose the one is that more convincing to the reader; not the one that is easier to perform
• Reviews that consider published papers are batter that those unpolished or those with thin evidence
• It has been said, with considerable truth, that most published research findings are false; and unpublished findings are worse.

​

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

13

Approaches to Measurement

• The purpose of experimentation is to take measurements that can be used as evidence
• Measurements can be quantitative, such as number or duration or volume_ the speed of a system, say, or an algorithm’s efficiency relative to a baseline.
• Measurements con also qualitative, such as an occurrence or difference_ whether an outcome was achieved, or whether particular features were observed
• From research question or hypothesis:
• What is to be measured?
• What measures will be used?
• Examine an algorithm
• Is it ok to measure execution time? If yes, what mechanism should be used to measure execution time?

​

Writing for Computer Science

Introduction

Computer Science

Information Systems

14

• What is measured should be connected to the aims of the research
• Research aims are qualitative
• We seek to improve an interface, accelerate an algorithm, Extract information from an image, generate better timetables for lectures, and so on
• Measurement is quantitative
• Find a property that can be represented as a quantity or value

Approaches to Measurement (continued)

Writing for Computer Science

Introduction

Computer Science

Information Systems

15

Good and Bad Science

• It is easier to distinguish valuable research from weak or pointless research
• Combination of formal and experiment based evidence is vital for good research
• Analysis cannot answer many questions immediately
• Theory without practical confirmation is not interesting
• Theoretical work should provide testable theories and thus must be tested
• Over- inflated terminology should be avoided:
• What is the difference between “hyperenet” and “network”?
• What do we understand by the words “intelligent”, “belief”, “aware”

Writing for Computer Science

Introduction

Computer Science

Information Systems

16

�

​

Reflections on Research�

​

• Two philosophers are arguing in a coffee shop. The coffee shop person goes over to them and asks:
• “What are you arguing about?
• “Were debating whether computer science is a science”, answers one of them.
• We’re not sure yet,” says the other. “we can’t agree on what ‘is’ means”.
• If the arguments and experiments are sound, if the theory can withstand skeptical scrutiny, if the work was undertaken within a firework of past research and provides a basis for further discovery, then it is science. Much computer science has this form.

Writing for Computer Science

Introduction

Computer Science

Information Systems

17

• By doing experiments, we cannot prove a hypothesis to be true
• Falsification: if experiment result in negative, We don't say that the hypothesis is false but say that some assumptions are wrong
• Thus, just redesign the experiment
• Confirmation: Confirm theories through experiments
• There is should be some experiment that proves the hypothesis is simply hypothesis to be wrong; else, hypothesis is simply uninteresting
• Ex: hypothesis: “a search engine can find interesting web page in response to queries”

Reflections on Research (continued)

Writing for Computer Science

Introduction

Computer Science

Information Systems

18