> ## Documentation Index
> Fetch the complete documentation index at: https://methodology.china-index.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Indicator development
> China Index methodology, 2021 edition, as published.
export const GradingMatrix = ({withNone = false}) => {
const rows = [...withNone ? [{
scale: 'None',
scores: [0, 0]
}] : [], {
scale: 'Few',
scores: [1, 3]
}, {
scale: 'More than a few',
scores: [2, 4]
}];
return
|
Significance →
Scale ↓
|
Insignificant |
Significant |
{rows.map(row =>
| {row.scale} |
{row.scores.map((score, i) => {score} | )}
)}
;
};
## Model of Influence
The indicators fall into three categories to measure PRC influence: exposure, pressure, and effect.
Most of the indicators ask about the exposure of a country to a mechanism of PRC influence - e.g. economic dependence, collaboration, or receiving some form of benefit. Some indicators ask about pressure overtly applied via such a mechanism - e.g. threatening economic punishment to provoke or prevent a political decision. Other indicators ask about the observable effects following from mechanisms of influence, or decision-making changes toward favor or align with PRC’s interests - e.g. allowing Huawei to build 5G networks. These three classes of indicators can be seen to form a causal chain, from exposure via pressure to effects, which appears sufficient to capture PRC influence.
It has been argued that the presence of exposure to mechanisms of PRC influence is insufficient to measure PRC influence, and therefore that such a measure should be restricted to consider effects. It has also been argued that the exposures are relevant, and of broader interest to potential consumers of the data we collect. We argue that if the final questionnaire contains both, then countries that experience effects will be correctly ranked as more influenced than those that do not, while capturing all the phenomena of interest.
## Conflicts or resistance
One complication we considered was that, after the application of overt pressure, it is possible that a country would respond not with submission leading to cooperation, but rather with resistance leading to conflict. We suggest dropping consideration of indicators that ask about the conflict to promote simplicity and conceptual clarity. Although such conflict can be seen to be closely related to PRC influence and fits into our model, it does not itself measure influence, but rather resistance to influence. Furthermore, in many instances, questions about resistance can be reworded to ask instead about submission - e.g., instead of asking whether a country has introduced legislation to ban Huawei from 5G networks, we can ask whether they have been included.
## How are the indicators generated?
Doublethink Lab facilitates and coordinates with the Questionnaire Committee to develop and select the most relevant questions for each domain throughout January and February 2021. The drafted questionnaire is now open to comments and will be finalized before March 15th, 2021.
## Design guidelines
The questionnaire will be designed with the following factors:
* **Survey length:** Each domain will contain 11 questions. A total of 99 questions will be sent to regional partners to conduct the research. Each question shall be designed to be answerable in 30 minutes of research.
* **Domain indicators:** Within each domain, each of the eleven questions will target a single observable feature of PRC influence that is both relevant and important to the scope of the domain. Each question will correspond to one indicator.
* **Question/indicator independence:** Each question will be independent of the others in the same domain. There will be no overlap between questions to avoid double counting.
* **Response recency:** Each question will focus on a single, precise issue that reflects the current status of PRC influence in the target country, focusing on events or incidents of the past 12 to 18 months (the calendar year of 2020 to the answering date).
* **Option scales:** Depending on the nature of the question, the options could either be designed for binary (yes or no) or graded from 0 to 4. Higher scale values will always indicate a higher degree of PRC influence.
* **Neutrality:** No question will invite value judgment on the presence or absence of PRC influence. Each question will focus on a single issue without subjective feeling or preference.
* **Balance:** Questions within a single domain will be designed to cover equal scope within that domain, such that each question/indicator merits equal weight within its domain.
## Indicator Development
Candidate indicators relevant to measuring PRC influence in each domain are suggested by the questionnaire committee in collaboration with Doublethink Lab. The candidate indicators are then modified to suit the criteria defining good indicators and to fit the grading system. Finally, a set of eleven indicators is chosen that capture the most important aspects of PRC influence.
Indicators are one of two types: binary (YES/NO) or graded. Binary indicators ask for the presence or absence of a particular phenomenon. Graded questions consider the scale and significance of the phenomenon (see next section).
## Grading
The goal of the grading system is to capture more nuanced data for further comparison between countries. The grading system addresses the problem that the scale and significance of an indicator can vary across countries, and therefore so does the level of PRC influence. For example, one indicator asks whether Chinese Students and Scholars’ Associations (CSSAs) are present in the country. In terms of scale, there may be a single CSSA or every university could have one. In terms of significance, the universities at which CSSAs are present could be among the most nationally important, or those with a negligible impact on social and political life.
We, therefore, provide a grading matrix that accounts for these two factors, mapping them to a score in the range of integers from one to four:
As there may be more than one CSSA, the question of significance is considered to be the maximum significance level - in other words, the most significant university at which there is a CSSA.
Significance and scale are high-level, coarse-grained concepts. For particular questions, they need to be given concrete interpretations. We suggest allowing regional partners and local experts to make these judgments.
The grading part of each indicator therefore becomes: 0 = no; 1 = few, but insignificant; 2 = more than a few, but insignificant; 3 = few, but significant; 4 = more than a few, and significant.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.