15 The quality and usefulness of data collected at the local level differ widely. Data quality may be limited by a number of factors. While some township officials express confidence in the quality of the data their departments collect, others are less confident, viewing it as “good enough” to satisfy the demands of superiors and give them a rough idea of the situation, but with no further concern for how it might be used. There are several obstacles to collecting high-quality data at the local level: z Lack of training. Primary data collection is usually carried out using paper forms by staff without special training on data collection. If a form is unclear, entries may be omitted or entered as written comments instead of values. Sometimes whole sections of a collection form are dispensed with and replaced with a written narrative on the back. It is not uncommon for overburdened government staff to have family members step in to help finish the job. Reassignments within departments are common, so staff tasked with collecting data one year may do something completely different the next, requiring new staff, with limited training, to take on the responsibility. z Lack of quality assurance. One significant obstacle to improving data quality—and the perception of quality—is the lack of quality assurance by local department officials. Proper data collection depends on following a careful plan. Forms should be clear in design, written in language the data collectors understand, with clear indications of the units of measurement to be used. In cases where data requests come from parent agencies, it is the parent agency’s responsibility to understand the local context and establish appropriate standards for data collection and mechanisms for quality assurance. It is unclear what form of quality assurance, if any, is routinely practiced. z Lack of ownership of collected data. Local officials are often uninvolved in the design of data-collection projects that they are called upon to manage, and the data requested by parent agencies tends to serve the requesting agency’s own needs. Without a clear explanation of how the data will be used, township officials can’t understand why it is being collected, or why in this form. Instead, it’s perceived as just another task to satisfy higher-ups, or as a record-keeping exercise rather than a crucial tool for decision-making. In the worst case, township officials may suspect that the data will be used as performance metrics, and they may feel pressure to alter the results. 2.3 HOW IS DATA STORED AT THE LOCAL LEVEL? Local-level data storage in Myanmar remains predominantly paper based. Data is often collected on paper forms, and township department offices often house significant stores of paper forms and ledgers, with manual filing mechanisms to support data retrieval. Duplicate copies may be made and shared with district, state/region, and Union offices. Paper storage has the benefit of simplicity, but paper data takes up a lot of space, can be damaged by fire or water, among other things, and is more difficult to copy than digital data. Where data is stored digitally, storage formats help facilitate aggregation and dissemination rather than analysis. In some departments, officials are tasked with converting paper-based data to digital. Where this is done, paper forms are often converted into computer files in formats like PDF or Word. These formats can help officials to aggregate data (i.e., combine data from multiple forms into a single data file), and they help officials share data more easily with district, state/ region, and Union offices, publish data on the internet, or print it and share it with other parties. As detailed below, however, these file formats do not permit easy data analysis. Digital data storage and recording standards have not yet been established in many department offices. While data is increasingly stored digitally in township government offices, many offices have no established standards for storing or sharing data—standards such as directory structures, folder- and file-naming conventions, and acceptable file formats. Instead, ad hoc strategies prevail, which makes data retrieval and sharing more challenging. Likewise, data-recording conventions have yet to be developed, which creates difficulties in understanding and analyzing data. For example, as shown in table A, data stored by government in Excel spreadsheets may contain blank entries (purple cells). Here it is not clear whether these entries should have been zero, they have yet to be tabulated, the data collection is incomplete, or the data has been withheld for another reason. This limits the usability of the data.